Computed tomography and enamel thickness of maxillary molars of Plio-Pleistocene hominids from Sterkfontein Swartkrans and Kromdraai (South Africa) An exploratory study.код для вставкиСкачать
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 89133-143 (1992) Computed Tomography and Enamel Thickness of Maxillary Molars of Plio-Pleistocene Hominids From Sterkfontein, Swartkrans, and Kromdraai (South Africa): An Exploratory Study GABRIELE A. MACHO AND J. FRANCIS THACKEXAY Departnient of Palaeontology and Palaeoenuzronmental Studies, Transuaal Museum, Pretoria 0001, South Afrrca KEY WORDS CT-scanning, Enamel thickness, Enamel distribution, k'-'lro-k'"iistocene maxiiiary molars, South M i c a , Huiiiiriid ABSTRACT This paper is one in a series which explores the possibility of using the non-destructive CT technique to identify patterns in tooth enamel distribution and structure of hominid molars from Plio-Pleistocene sites in South Africa, notably Swartkrans, Sterkfontein, and Kromdraai. Whereas previous investigators have emphasised gross differences in absolute and relative or average enamel thickness between hominid taxa, the present study highlights differences in enamel thickness over functionally significant regions of the crown. Differences in the distribution of enamel in A. robustus, A. africanus, and Homo sp. are identified through the use of bivariate and multivariate analyses, and are interpreted in terms of dietary regimes. 0 1992 Wiley-Liss, Inc. Attention has recently been drawn to differences in tooth enamel thickness among Plio-Pleistocene hominids (Beynon and Wood, 1986; Conroy, 1991; Conroy and Vannier, 1991; Gantt, 1977, 1983; Grine and Martin, 1988; Kay, 1981;Martin, 1985; Wallace, 1972,1978). Several of these researchers have concluded that Ausfralopithecus robustus and A. boisei have "hyper-thick enamel, while A. africanus and Homo are characterized by relatively thinner enamel. Furthermore, these studies confirmed that enamel is not equally distributed over the entire tooth crown, but thicker in areas of greater functional demands (Molnar and Gantt, 1977; Molnar and Ward, 1977). Most of the studies on enamel thickness relied on measurements taken on naturally broken or worn teeth. Martin (1985:260) stressed that such measurements have "limited value." So far a total of only six Plio-Pleistocene hominid teeth from East and South African sites have been sectioned (Grine and Martin, 1988). Despite the greater accuracy of results offered from this destructive approach, it is at present unlikely that many more pre0 1992 WILEY-LISS, INC cious fossil teeth will be sacrificed for sectioning. It is therefore desirable to take maximum advantage of non-destructive techniques in order to determine variability in enamel thickness and distribution within and between taxa. Zonneveld and Wind (1985) pointed out the potential of computed tomography ICT'! for palaeoanthropology with special reference to measuring enamel thickness. From sagittal scans they recorded a maximum thickness of 2.6 mm in M3 of SK 48 (A.robustus from Swartkrans), and 3.3 mm in M2 of TM 1517 (A.robustus from Kromdraai). Recent research has been conducted to investigate further the possibility of using CTscans on isolated teeth. In Grine's (1991a,b) study of extant primates, measurements obtained from CT-scans were compared with those taken from SEM images of the same specimens. By contrast, Conroy (Conroy, 1991; Conroy and Vannier, 1991) took mea- Received May 28,1991; accepted March 10, 1992. G.A. MACHO AND J.F. THACKERAY 134 TABLE 1. List of horninid maxillary molars from Swartkrans, Sterkfontein, and Kromdraai included in this study iL = left. R = right) Cat. No. Side M’ M” M“ No. Site Saecies Sterkfontein Sterkfontein Sterkfontein Sterkfontein Sterkfontein Swartkrans Swartkrans Swartkrans Swartkrans Swartkrans Swartkrans Swartkrans Swartkrans Swartkrans Swartkrans Swartkrans Swartkrans Swartkrans Swartkrans Kromdraai Kromdraai Kromdraai Kromdraai 1 2 3 4 5 6 7 8 9 10 11 12 13 i4 15 16 17 18 19 20 21 22 23 ~ Sts Sts Sts Sts TM SK SK SK SK SK SK SKX SK SK SK SK SK SK SKW Kb Kb TM TM 8 24 28 37 1511 36 41 89 98 102 105 268 829 832 839 839 3977 3975 14129 5063 5383 1601 1603 L R R L R R X X X x x x x X X X R L L L R R L R R R L L x X X x X X X x x X X X X A. africanus A. ufricarzus A. africaizus A. africanus A. africanus A. robustus A. robustus A. robustics A. robustus A. robustus A. robustus Homo sp. A. robustus A. robustus A. robustus A. robustus A. robustus A. robustus A. robustus A. robustus A. robustus A, robustus A. robustus ~~ Catalogue specimen numbers correspond to specimen numbers 1-23 included in Figures 2 and 3 tocene hominids from South African sites were selected for analysis from collections housed at the Department of Palaeontology and Palaeo-environmental Studies, Transvaal Museum, Pretoria (Table 1).The hominid sample from Sterkfontein comprise specimens collected by Robert Broom and John Robinson between 1936 and 1958. Those from Swartkrans were obtained by Broom and Robinson, and more recently by C.K. Brain. The material from Kromdraai was collected by Broom, supplemented by isolated hominid teeth from Elisabeth Vrba’s excavations. The specimens from Sterkfontein derive from Member 4 and have been attributed to A. africanus, while all except one tooth from Swartkrans and all those from Kromdraai have been assigned to A. robustus. SKX 268 from Swartkrans has been described as Homo (Grine 1989). A Siemens SOMATOM DR3 scanner a t the Department of Radiology, Hillbrow Hospital, Johannesburg, was used. One millimeter slices (i.e., the thinnest possible slice width) were taken with a high resolution filter at 125 kV and a scanning time of 7 seconds. The images were reconstructed MATERIALS AND METHODS from a pixel matrix of 512’ and 720 projecIsolated unworn or slightly worn perma- tions. The window settings were kept connent maxillary molars M1-M3 of Plio-Pleis- stant for all images. surements directly from CT-scans of fossil hominid teeth. Whereas Conroy maintained that his study demonstrated the viability of the CT-scanning technique, Grine contended that it could not be used to provide reliable and accurate measures of enamel thickness. Measurement errors on CT scans stem from different sources and the percentage error may vary considerably (Koehler et al., 1979; Ruff and Leo, 1986; Zonneveld, 1983). While recognizing difficulties associated with the recording of measurements from hard copy outputs of CT-scans, we place emphasis on patterning in enamel thickness within teeth and between individuals, rather than on absolute values. In this investigation we make use of the non-destructive CT scanning technique (Computed Tomography) to explore variability in tooth enamel distribution of hominids represented at Swartkrans, Sterkfontein and Kromdraai in the Transvaal Province, South Africa; specimens from these sites have been commonly attributed to A. africanus, A. robustus, or Homo. PATTERNING IN ENAMEL THICKNESS PROTOCONE 135 PARACONE Fig. 1, Cross-section through maxillary molar showing seven linear dimensions of enamel thickness (variables 1-7) used in this exploratory study. Each measurement was scaled to the reference plane A-B being 10 mm. The basal area of each tooth was orientated horizontally and mounted on a microscope slide on a soft plasticine base. For purposes of orientation two very small, low density markers were placed on tips of the paracone (mesio-buccal cusp) and protocone (mesio-lingual cusp). The microscope slide with the mounted tooth was placed on another slide whose lateral edge was placed beneath the paracone-protocone plane. This arrangement facilitated accurate alignment of both cusp tips on the beam of the CTscanner. A l m m slice was then taken in the bucco-lingual plane, perpendicular to the basal area. In order to test for accuracy of orientation, additional scans were taken at lmm increments, mesially and distally from the plane initially selected. Bucco-lingual diameters. which had previously been recorded on original specimens, were checked against corresponding diameters of the images on the screen with the built-in device. CT images were recorded on film (hard copy) and the data stored on floppy discs. Measurements of external dimensions of each tooth were also taken from hard copies and checked against the same diameters recorded directly from originals. A further check of CT-imagery was made by examining the dentine horns below cusps of the protocone and paracone in a series of scans. Scans that had not been taken through both cusp tips simultaneously could be identified when the apical angle of the dentino-enamel junction (DEJ) was obtuse under one cusp contrasting with a more acute angle below the other; a scan further mesially or distally usually resulted in a slight improvement and an apparently more accurate delineation of one dentine horn but not of the other. In such cases the specimen could be re-orientated to correct for the observed disparity. Of the specimens included in this analysis (Table 11, comparisons were made between external distances recorded from the teeth and those from hard copies of CT images. A correlation coefficient (r2)of 0.99 was obtained. However, preliminary analyses suggest that small distances may be consistently overestimated; this would be in accordance with Grine's study of modern specimens (1991a,b). It is recognised that such overestimates for particular variables are not necessarily consistent between teeth. On the other hand, since errors from CT scans are likely to by systematic within any particular tooth, proportional distribution of measurements of enamel within the tooth should not be affected. All images were traced and enlarged before measurements were taken with a Kontron Image Analyser. Where necessary the outlines were adjusted for slight wear (specimens SK 41, SK 832, SK 3975, SK 3977, SKW 14129 and TM 1511). Seven variables of enamel thickness, including six previously defined by Beynon and Wood (1986) and Martin (19831, are shown in Figure 1. The data were analysed by means of commercially available statistical packages (Lotus 2.1; Statgraphics 4.0) and include bivariate plots, linear regression analyses, and principal component analyses. All teeth were scaled t o the same size, the distance A-B (Fig. 1)being standardised arbitrarily to 10 mm. 136 G.A. MACHO AND J.F. THACKEMY All measurements were taken by one of us (G.A.M.).A selected set of variables was recorded twice to check for consistency. Comparison between these sets of data yielded correlation coefficients (r2)ranging between 0.97 and 0.99. Although SK 839 is represented by antimeres, these teeth will be treated as separate entities. This was done not to artificially inflate sample size but rather as an attempt to judge the degree of variability occurring within one individual. RESULTS In Figure 2, lingual enamel diameters (variable numbers 1, 3, and 4) are plotted against corresponding enamel thickness on the buccal side (numbers 2,6, and 5, respectively). The results demonstrate a considerable degree of overlap between sites. Although this may be due to some extent to systematic measurement error, it is nonetheless evident that the highest values of enamel thickness are obtained for specimens of “robust” australopithecines. Deviations from the isometric line are apparent. Close inspection reveals that this trend is most strongly expressed for measurement No. 3 vs. 6 (Fig. 2). Least-squares linear regression analyses were performed between pairs of variables and using the pooled sample of hominid teeth. The enamel thicknesses of the lingual cusp of the tooth were chosen as the independent variables. Regression analyses between variables 3 and 6 yield a slope of 0.57. At the occlusal basin (variables 4 vs. 5) there is also a disparity between the enamel thicknesses of the lingual and buccal cusps; the slope is 0.72. By contrast, when comparing lingual and buccal enamel thicknesses at cusp tips (variables 1 and 21, the regression line more closely follows the isometric line; in this instance, the slope is 0.86. The proportional distribution of enamel thickness in the Homo specimen SKX 268 deserves special mention. This molar exhibits thick enamel along the occlusal surface (variables 1,2,4,and 5) comparable to those of most other Swartkrans specimens and is positioned closely to the isometric line (Fig. 2). Along the lingual and buccal wall of the crown (variables 3 and 61, SKX 268 shows I I5x93 S ’ 21 3- . a 1 - 1.o 2.0 3.0 4.0 VARIABLE 1 (LINGUAL) 3- 2- 1.o 2.0 3.0 4.0 VARIABLE 4 (LINGUAL) 3- 2- 1.o 2.0 3.0 4.0 VARIABLE 3 (LINGUAL) Fig. 2. Bi-variate plots of pairs of corresponding variables on lingual and buccal sides of maxillary molars from Swartkrans, Sterkfontein, and Kromdraai. Dimensions of all variables are expressed in arbitrary numbers. Note the deviation from the isometric line (dotted). Specimens 1-5: Sterkfontein; specimens 6-19: Swartkrans; specimens 20-23: Kromdraai. thinner enamel than most of the other teeth from Swartkrans. Furthermore, it is the only specimen which clearly lies above the isometric line for variables 3 and 6. This indicates that the enamel is relatively thicker along the buccal wall in proportion t o the 137 PATTERNING IN ENAMEL THICKNESS TABLE 2 Principal component analysis, agenualues and percentage contributtons of components I-IV to the totol Measurement No. _ _ --. --_ 1 2 3 4 5 6 7 8 variance . Description . . .. ._ .___Protocone tip Paracone tip Protocone lingual Protocone buccal Paracone lingual Paracone buccal Intercuspal fissure iinrinnr~ PCI - - ____ __ __ ______ - _0 45 0 45 0 47 0 31 0 29 0 32 0 30 78 41 lingual one, whereas the opposite holds for other teeth studied. Principal Components Analysis (PCA) was employed in order to explore structural differences between specimens (Table 2, Fig. 3). The first principal component (PC) is associated with overall size differences, and accounts for 78% of the variance. The contributions of PC I1 to PCIV are relatively small, but nonetheless informative since they discriminate between shapes and patterns of enamel distribution in individual teeth. PC I1 contrasts enamel thickness on the cusp tip (variable 2) and occlusal surface (variable 5) of the paracone, with enamel thickness on the lingual slope of the protocone (variable 3). This PC clearly separates Homo SKX 268 from other specimens (Fig. 3). The third PC reveals general differences between sites: all specimens from Sterkfontein have negative coefficients: Kromdraai specimens Kb 5383, TM 1601, and TM 1603 also fall in that category, as do SK 3977, SK 36, and the antimeres of SK 839, but all other Swartkrans teeth (n = 10) and Kb 5063 have positive coefficients (Fig. 3). The contrasting features of PC 111 are the dimensions of the buccal side of the paracone (variables 2 and 6) and that of the lingual wall of the protocone (variable 3), on one hand, and the remaining variables of enamel thickness on the occlusal surface and the protocone, on the other (Table 2 ) . It would thus seem that there is a difference in relative enamel thickness between the lingual and buccal side of hominid teeth from different sites, which is not detected by bi-variate analyses. The fourth PC contrasts enamel thickness of the fissure between the protocone and paracone (variable 7) with thickness of vari- Principal components PCII - PCIII __ PCN 0.07 -0.62 0.71 -0.09 -0.31 0.12 0.01 9.07 -0 41 -0 0 3 -0 01 -0 31 0 06 0 11 0 84 4 58 0.23 -0.52 -0.29 0.64 0.15 -0.23 0.34 5.20 -2 -l PC II -2 -l PC II -11 -2 ~ , , I ~ -1 , 1 O -1 0 1 I 1 PC I1 Fig. 3. Plot of principal components 1 and I1,II and 111, and I1 and IV. Specimens 1-5: Sterkfontein; specimens 6-19: Swartkrans; specimens 20-23: Kromdraai. 138 G.A. MACHO AND J.P THACKEIUY ables 1 and 4 on the buccal slope of the protocone. Inspection of the plot reveals that this component separates three of the Kromdraai teeth from other Specimens. TM 1603 is most clearly separated, whereas Kb 5063 and TM 1601 lie adjacent to the main cluster. The remaining principal components (PC V, VI, and VII) account for less than 2% of the total variation, and do not reveal clearcut patterns. Only PC VI appears to differentiate between MI, M2, and M3. DISCUSSION It is clear that valuable information can be obtained from CT-scans provided certain guidelines are followed (see also Ruff and Leo, 1986; Spoor, pers. comm.). In this study we have reported a correlation coefficient (r2) of 0.99 for the relationship between external measurements taken from original specimens and those obtained from hard copy outputs of CT images. Difficulties of course arise when examining internal parameters. Although window settings were kept constant in the present study, differences in CT images may occur on account of preservational and other factors (depending on whether roots, bone, matrix or plaster are attached to the crowns). Here we have refrained from relying on the absolute values of CT based measurements of enamel thickness, and instead investigated proportions and patterns of enamel distribution wit,hin individual teeth. Thin, finely tapered cervical margins of the enamel cap form blunt borders on CTscan images and are situated towards the pulp cavity, especially in cases of isolated crowns. Measurements of areas and volumes would be especially affected by this fact (Spoor and Zonneveld, pers. comm.). Hence, only linear dimensions were recorded in the present study. These diameters are close to the occlusal part where enamel thickness is expected to be relatively uniform (Grine and Martin, 1988). However, distortion of the DEJ a t the cervix may have also affected the position of the reference plane A-B, which is parallel to the cervical border, but the edges on both sides are affected by beam hardening and edge arte- facts t o a similar extent. Thus, inclination of the reference plane from the true plane may be regarded as slight if not negligible. Damage of the DEJ appears to be a potential cause of error which is difficult to assess on the scans. Irregularities of the DEJ in CT-scans are the only indication of possible breakage a t the DEJ. Although the specimens used in the present study all appear to have an intact DEJ, the possibility of damage cannot be ruled out completely, Whereas it is technically possible to obtain accurate rstimates of enamel thickness fi.~niCTscans (Spoor, pers. comm.), it is questionable whether the technique will advance to such an extent that morphological details of the border between the two tissues can be accurately determined. This must be regarded as one of the major drawbacks of this new, non-destructive technique, which offers such promising prospects in other respects (e.g., Ruff, 1989; Vannier and Conroy, 1989). Differing enamel thicknesses within and between teeth are usually regarded as evolutionary responses to functional demands through dietary specializations (e.g., Kay, 1981; Molnar and Gantt, 1977; Ward and Pilbeam, 1983). Tooth morphology suggests that the “robust” australopithecines were adapted to a coarser diet than that of “gracile” specimens (Grine, 1981, 1986; Jolly, 1970; Kay, 1975; Robinson, 1954, 1963). Furthermore, A. robustus and A. boisei have been described as “hyper-thick enamelled hominids (Beynon and Wood, 1986; Grine and Martin, 1988), while H. sapiens and A. africanus have proportionally thinner enamel. This condition persists even when the enamel is scaled to tooth size. Although Robinson (1956) acknowledged some differences in thickness of the enamel cap between australopithecine species, he claimed that the differences are not substantial. Overall, the present study shows a considerable degree of overlap in enamel thickness of “gracile”and “robust” australopithecines. A. robustus, as represented by specimens from Swartkrans, tends to be associated with high values (Figs. 2, 3). Although our measurements of specimens in this study are comparable t o those of published enamel PATTERNING IN EN AMEI, THICKNESS thicknesses of Plio-Pleistocenehominids, we shall refrain from placing too much emphasis on absolute diameters until we have experimentally evaluated the effects of beam hardening and beam attenuation in enamel and appraised the absolute measurement error. More importantly, distinct differences in the distribution of enamel have been observed within teeth, and between teeth belonging to different taxa. Inspection of the bi-variate plots (Fig. 2) reveals that patterns of enamel distribution are apparer,t?jr c o r n ~ ~ oton A. clfricanxs (Specimens from Sterkfontein Member 4) and A. robustus (represented by samples from Swartkrans and Kromdraai). Furthermore, on the basis of samples included in this study of enamel thickness, there is no strong evidence to suggest that there is necessarily more than one species among the Sterkfontein Member 4 specimens, as suggested by Clarke (1988a,b). In all of the teeth examined in this study, with the exception of the Homo molar SKX 268, the lingual wall of the protocone exhibits thicker enamel than the buccal slope of the paracone, whereas the buccal slope of the protocone and the lingual slope of the paracone are more similar in enamel thickness. Enamel thickness at cusp tips exhibits the widest scatter in all of the bivariate plots shown in Figure 2. This finding could be attributed, at least in part, to the fact that adjustment had been made for slight wear in some specimens. Nonetheless, a difference in enamel thickness has been found between cusp tips at the protocone and the paracone. Generally (with the exception of TM 16031, lingual cusps exhibit thicker enamel. Similar differences have also been found by Beynon and Wood (1986) in east African “robust” specimens, and by Grine and Martin (1988) in molars of modern humans, A. africanus and A. robustus (P. crassidens). Such differences reflect greater functional demands on the lingual than on the buccal cusps in maxillary molars, whereas the opposite holds for mandibular molars (Molnar and Ward, 1977; Molnar and Gantt, 1977). It is thus surprising that Conroy (1991)lists only 9 out of 21 lower hominid molars from South Africa to possess thicker enamel on 139 the buccal than on the lingual side, while the present study, also on South African hominids, accords with the expected pattern. This discrepancy could be due to differences in technique which need to be investigated and clarified. Bearing in mind the reported functional demand on the buccal and lingual cusps (Molnar and Gantt, 1977; Molnar and Ward, 1977), it is surprising that the greatest deviation from the isometric line was found for tooth walls rather than cusp tips. Although this could xiggect that the entire lingual cusp was under relatively stronger compressive forces, a note of caution seems to be called for. Australopithecines are characterized by a high percentage of protoconal cinguli or Carabelli cusps, especially A. africanus (Robinson, 1956; Sperber, 1974; Wood and Engleman, 1988). Besides, in a sample of Homo s. sapiens from Sangiran, Korenhof (1960,1982) showed a relatively high occurrence of complete or incomplete cinguli a t the DEJ despite the lack thereof at the tooth crown. A similar morphological situation at the DEJ may have exaggerated differences between lingual and buccal walls of teeth used in the present study. With regard to enamel thickness at the occlusal basin, SKX 268 (Homo)falls within the upper range of values, whereas the buccal and lingual walls of the tooth exhibit relatively thinner enamel when compared to values from most other specimens. Hence, in total enamel aredvolume of enamel SKX 268 could be expected to lie at the lower end of the range for Swartkrans hominid teeth. While bi-variate analyses appear to show no major structural differences in enamel thickness distribution among australopithecine specimens from Sterkfontein, Swartkrans and Kromdraai, multivariate analyses reveal complex patterns of enamel distribution over the tooth crown between teeth. Most of the variation between maxillary molars can be attributed to absolute size differences, although this could have been artificially inflated because of measurement errors. Complex proportional differencesin enamel thickness of sectioned Plio-Pleistocene hominid molars were noted by Grine and Martin 140 G A MACHO AND J.F. TKACKERAY (1988). The A. africanus specimen STW 284 and A. robustus SKX 21841 exhibit a pattern similar to that found in humans, whereas the other A. africanus tooth, Stw 402, is comparable to A. boisei L10-21 and L398847. Kb 5223 is similar to the latter in its preserved parts. Although Grine and Martin seem to acknowledge that humans tend to be different in their proportions of enamel thickness from the “robust” australopithecines, it is not clear whether they regard the results obtained for A . africanus as random nwing t o the qmall sample si7e, n r whether their findings have taxonomic implications (sensu Clarke). The Homo specimen SKX 268 was most clearly separated from the other teeth, mainly due to a proportional difference of enamel thickness between the lingual half of the paracone and that of the lingual slope of the protocone (Table 2, Figs. 2,3);the former was proportionally thicker than the latter. Two explanations, not mutually exclusive, are offered for this finding. The lingual slope of the paracone is a Phase I facet in primates and thus participates mainly in shearing (Crompton and Hiiemae, 1969; Kay, 1975; Maier and Schneck, 1981, 1982; Mills, 1955, 1963). A more omnivorous diet in Homo would have required proportionally more puncture-crushing and shearing than grinding, the latter being essential to break down harder and more fibrous foods. The morphology of Homo teeth with high, less spaced cusps and more steeply inclined slopes of wear facets is a reflection of an altered diet. Hence, a proportionally thickened lingual slope of the paracone relative to the tooth walls could be regarded as an evolutionary adaptation to enhanced shearing and less grinding activity in Homo. Although the cusps of australopithecine molars are low and bulbous (Beynon and Wood, 1986; Grine, 1981, 1986; Robinson, 1956; Wallace, 1972, 19781, those of Homo are more pronounced. This situation may be paralleled at the DEJ. No systematic study has so far been done on this aspect, but Grine and Martin’s figure 1.10 and 1.11 clearly support this suggestion. Owing to the more protruding dentine horns in Homo, 1 the reference plane A-B (Fig. 1)is proportionally lower and thus closer to the cervical border where the enamel thickness tapers. It is more likely that in australopithecine teeth, appositional, i.e. full-thickness enamel over the cusp tips, rather than imbricational enamel has been measured (Beynon and Wood, 1987). In Homo the enamel extension rate decreases markedly towards the cervix, while the crown formation time is longer when compared with australopithecines (Beynon and Dean, 1988; see also Mann et a1 , 19W for review nf literatvrcl.) This fa.ct as well as the frequent occurrence of protoconal cinguli in Australopithecus discussed above may have exaggerated (although not invalidated) structural differences between Homo and australopithecine specimens examined. The use of this measurement for taxonomic purposes is therefore questionable. It would seem that diameters of walls defined by Beynon and Wood (1986) (i.e., their variables LT(L) and LT(B)),which are taken about lmm below the dentine horns irrespective of the reference plane A-B, are more reliable than those of Martin (1983) (i.e., his variables k and 1; see also our measurements No. 3 and 6). The latter pair could be abandoned for taxonomic purposes. Grine and Martin (1988) did not report a distinctive pattern of enamel distribution for their sample from Sterkfontein. However, the present study demonstrates that the Sterkfontein material differs structut ally from the majority of Swartkrans molars, but not from those from Kromdraai. Sterkfontein was separated from the main cluster of Swartkrans hominid teeth by a proportional difference of enamel thickness between the occlusal basin (including the cusp area of the protocone) on the one hand, and the buccal half of the paracone and the lingual wall of the protocone, on the other. Interestingly, three out of four Kromdraai specimens, i.e., Kb 5383, TM 1601, and TM 1603, fell on the same half of PC I11 as did four Swartkrans specimens. Hence, most of the molars studied from Swartkrans have a proportionally thicker occlusal basin when compared with those from Sterkfontein and Krorndraai. PATTERNING IN ENAMEL THICKNESS In his SEM study on “gracile”and “robust” deciduous australopithecine teeth, Grine (1981) claimed that Kromdraai molars appeared to have followed a masticatory occlusal course similar to that of A. africanus. He concluded that on morphological grounds the Kromdraai hominid molars seem “to be ‘intermediate’ between the ‘gracile’ and Swartkrans ‘robust’ forms” (Grine, 1981: 223).The pattern of enamel thickness distribution, as found in the present study, substantiates this claim. While there are subtle differences between the majority of specimens from Sterkfontein and those from Kromdraai, the hominid teeth from these sites are more similar to each other than they are to most of the specimens of A. robustus from Swartkrans with regard to enamel distribution. However, the Kromdraai specimens investigated are not homogeneous morphologically; i.e., none of the principal components resulted in a complete clustering of all hominid teeth from that site. The proportionally thicker buccal slope of the protocone can be reconciled with Grine’s (1981, 1986) findings on tooth wear in A. robustus. He suggested that the greater Phase I1 facets in “robust”australopithecine molars are related to a different diet and a greater amount of compressive force. The morphology of the teeth and the microwear features both led him to conclude that the dentition of A. robustus was adapted for grinding and puncture-crushing, The distribution of enamel found in the present study supports this hypothesis. Although the specimens from Swartkrans have on average thicker enamel, the major difference seems to occur along the Phase I1 facets; the areas underneath these facets are endowed with proportionally thicker enamel. This could indicate an evolutionary response to enhanced grinding activity attributable to fibrous diet. Conversely, A. africanus, which possesses more pointed cusps, also exhibits proportionally thinner enamel at the grinding facet of the protocone. The distribution of enamel in A. robustus from Kromdraai is comparable to that seen in maxillary molars from Sterkfontein. In conclusion, structural differences appear to accord, at least in part, with differ- 141 ences between taxa and may be related to different diets. However, since A. africanus and A. robustus have in common relatively thick enamel along the lingual and buccal wall, a substantial compressive load owing to puncture-crushing and grinding can be postulated for both. By contrast, Homo appears to exhibit a greater disparity in enamel thickness between the occlusal surface, including cusp tips, and walls o f the molar. In order to remain functional over a long period of time, a tooth so endowed should not be subjected to n;assive grinding forces. Nonetheless, even an omnivorous diet could demand a considerable amount of shearing and puncture-crushing and, hence, thick enamel over cusp tips could be an advantage. However, since only one Homo molar has been used in this investigation, the results must be regarded as tentative. From this preliminary study we conclude that enamel thickness per se, or the volume of tissue, may not necessarily be a useful measure to distinguish between hominid species as postulated by previous anthropologists. Among early hominids, there appears to be considerable overlap in enamel thickness and overall tooth size and morphology, especially where maxillary molars are concerned (Corruccini and McHenry, 1980; Wood and Engleman, 1988). On the other hand, the distribution of enamel thickness over the tooth crown differs between taxa in a way that may reflect masticatory differences, indicative of different diets. The present results support conclusions reached from previous studies, which postulate that diets of “robust” australopithecines were comparatively coarser and more fibrous than those ofA. africanus and Homo. The possibility of enamel patterning reflecting evolutionary/functional responses to different diets needs to be tested further by utilizing a bigger sample of molars from extinct taxa and extant species with known dietary specialization. Furthermore, differences of enamel distribution between molars within a species need to be investigated. Finally, in order to conclusively elucidate functional and mechanical properties of enamel thickness, other aspects of tooth morphology and histology need to be incor- 142 G.A. MACHO AND J.F. THACKERAY porated. Such studies would preferably combine investigations of tooth morphology, development, microwear, enamel thickness, trace elements and stable carbon isotope ratios. ACKNOWLEDGMENTS We thank J . Nach for permission to use the Siemens Scanner at the Hillbrow Hospital, Johannesburg, and R. Harman and R. van der Riet for their help with the scanning operation. Furthermore, we are gratef‘d tc G. Conrog, M.C. Dean. L. Freedman, F. Grine, D. Panagos, E.E. Sarmiento, F. Spoor, and F. Zonneveld for valuable advice andlor comments on the manuscript. We also thank two anonymous referees for their thoughtful comments. R. Hill gave permission to use the Kontron Image Analyser at the Department of Anatomy, University of the Witwatersrand. This work was financed in part by the Foundation for Research and Development. We thank C.K. Brain, Director of the Transvaal Museum, for his support and encouragement. LITERATURE CITED Beynon AD, and Dean MC (1988) Distinct dental development patterns in early fossil hominids. Nature 335r509-514. Beynon AD, and Wood BA (1986) Variations in enamel thickness and structure in East African hominids. Am. J. Phys. Anthropol. 70:177-193. Beynon AD, and Wood BA (1987: Pattern. and rates o f enamel growth in the molar teeth of early hominids. Nature 326:493-496. Clarke RJ (1988a) A new AustruEopithecus cranium from Sterkfontein and its bearing on the ancestry of Purunthropus. In FE Grine (ed.j: Evolutionary History of the “Robust” Australopithecines. New York: Aldine de Gruyter, pp. 285-292. Clarke RJ (198813) Habiline handaxes and Paranthropine pedigree a t Sterkfontein. World Archaeology 2O:l-12. Conroy GC (1991) Enamel thickness in South African Australopithecines: Noninvasive determination by computed tomography. Palaeont. afr. 28:53-59. Conroy GC, and Vannier MW (1991) Non invasive evaluation of enamel thickness and volume in South African Australopithecines by computed tomography. Am. J . Phys. Anthropol., Suppl. 12t60 (Abstract). Corruccini RS, and McIlenry HM (1980) Cladometric analysis of Pliocene hominids. J. Hum. Evol. 9t209221. Crompton AW, and Hiiemae K (1969) Functional occlusion in Tribosphenic molars. Nature 222r678-679. Gantt DG (1977) Enamel thickness and human evolution. Am. J. Phys. Anthropol. 66r171. Gantt DG 11983) The enamel of Neogene hominoids. Structural and phyletic implications. In RL Ciochon and RS Corruccini (eds.): New Interpretations of Ape and Human Ancestry. New York: Plenum Press, pp. 249-298. Grine FE (1 981) Trophic differences between ‘gracile’ and ‘robust’ Australopithecines: A scanning electron microscope analysis of occlusal events. S. Afr. J. Sci. 77t203-230. Grine FE (1986) Dental evidence for dietary differences in Airtrolnnith~ciiaand Pnmnfhrnp7rs. a qnnntitativr analysis o f permanent molar microwear. J . Hum. Evol. 15t783-822. Grine FE (1989) New hominid fossils from the Swartkrans formation (1979-1986 Excavations). Craniodental specimens. Am. J . Phys. Anthropol. 79:409449. Grine FE (1991a) Use of computed tomography to measure tooth enamel thickness. Am. J . Phys. Anthropol., Suppl. 12:84 (Abstract). Grine FE (1991b) Computed tomography and the measurement of enamel thickness in extant hominoid primates: Implications for its palaeontological application. Palaeont. afr. 28t61-69. Grine FE, and Martin LB (1988) Enamel thickness and development in Austrulopithecus and Purunthropus. In FE Grine (ed.): Evolutionary History of the “Robust” Australopithecines. New York Aldine de Gruyter, pp. 3-42. Jolly CJ (1970) The seed-eaters: A new model of Hominid differentiation based on a baboon analogy. Man 5:5-26. Kay RF (1975) The functional adaptations of primate molar teeth. Am. J . Phys. Anthropol. 43t195-216. Kay RF (1981) The nut-crackers-A new theory of the adaptations of the Ramapithecinae. Am. J. Phys. Anthropol. 55:141-151. Koehler PR, Anderson RE, and Baxter B 11979) The effect of computed tomography viewer controls on anatomical measurements. Radiology 130:189-194. Korenhof CAW (1960) Morphological aspects of the human upper molar. Utrecht:Uitgeversmaatschappij Neerlandia. Korenhof CAW (1982) Evolutionary trends of the inner enamel anatomy of deciduous molars from Sangiran (Java, Indonesia). In R Kurten (ed.): Teeth: Form, Function and Evolution. New York: Columbia University Press, pp. 350-365. Maier W, and Schneck G (1981) Konstruktionsmorphologische Untersuchungen am GebiR der hominoiden Primaten. Z. Morphol. Anthropol. 72:127-169. Maier W, and Schneck G (1982) Functional morphology of hominoid dentitions. J . Hum. Evol. 11:693-696. Mann A, Lamp1 M. and Monge J (1990) Patterns of ontogeny in human evolution: evidence from dental development. Yrbk. Phys. Anthropol. 33:111-150. Martin LB (1983) The relationships o f the later Miocene PATTERNING IN ENAMEL THICKNESS Hominoidea. Ph.D. thesis, University of London. (as quoted by FE Grine and LB Martin, 1988). Martin L (1985) Significance of enamel thickness in hominoid evolution. Nature 314:260-263. Mills JRE (1955) Ideal dental occlusion in the primates. Dent. Pract. 6t47-61. Mills JRE (1963) Occlusion and malocclusion of the teeth of primates. In DR Brothwell (ed.): Dental Anthropology. Oxford: Pergamon Press, pp. 29-51. Molnar S, and Gantt DG (1977) Functional implications of primate enamel thickness. Am. J. Phys. Anthropol. 46t447-454. Molnar S,and Ward SC (1977) On the Hominid masticatory complex: Biomechanical and evolutionary perspectives. ,J. Hum. Evol. 6.551-568. Robinson J T (1954) Prehominid dentition and hominid evolution. Evolution 8:324-334. Robinson J T (1956)The dentition of the Australopithecinae. Transv. Mus. Mem. Ytl-179. Robinson J T (1963) Adaptive radiation in the Australopithecines and the origin of man. In FC Howell and F Bourliere (eds.): African Ecology and Human Evolution. Chicago: Aldine, pp. 385-416. Ruff CB, and Leo F P (1986) Use of computed tomography in skeletal structure research. Yrbk. Phys. Anthropol. 29t181-196. Ruff CB (1989) New approaches to structural evolution of limb bones in primates. Folia Primatol. 53t142-159. 143 Sperber G (1974) Morphology of the cheek teeth of early South African hominids. Ph.D. thesis, University of the Witwatersrand. Vannier MW, and Conroy GC (1989)Three-dimensional surface reconstruction software system for IBM personal computers. Folia Primatol. 53:22-32. Wallace JA (1972) The dentition of the South African early hominids: a study of form and function. Ph.D. thesis, University of the Witwatersrand. Wallace J (1 978) Evolutionary trends in the early hominid dentition. In CJ Jolly (ed.): Early Hominids of Africa. London: Duckworth & Co. Ltd., pp. 285-310. Ward SC, and Pilbeam DR (1983) Maxillofacial morphology of miocene hominoids from Africa and IndoPakistan. In RL Ciochon and RS Corruccini (eds.): New Interpretations of Ape and Human Ancestry. New York: Plenum Press, pp. 211-238. Wood BA, and Engleman CA (1988)Analysis of the dental morphology of Plio-Pleistocene hominids. V. Maxillary postcanine tooth morphology. J . Anat. 161:1-35. Zonneveld FW (1983) Computed Tomography. The Netherlands: Philips Publication. Zonneveld FW, and Wind J (1985) High-resolution computed tomography of fossil hominid skulls: A new method and some results. In PV Tobias (ed.): Hominid Evolution: Past, Present and Future. New York: Alan It. Liss, pp. 427436.