Materials and Methods

Available replicas produced from molds used by Ungar et al. (2006b) were examined using white-light confocal microscopy. Eighteen specimens in total were studied (Table 11.3). This sample included most of the same individuals analyzed by Ungar et al. (2006b). Three specimens from that study (OH 7, Stw 82, and SK 2635) were unavailable for analysis, but three others were added (KNM-BK 8518, KNM-ER 992, and KNM-ER 1808) with the detection of previously unidentified antemortem microwear using the confocal microscope. Any study with limited sample sizes can be affected by the exclusion of a few specimens and inclusion of others, so comparisons between results presented here and those in Ungar et al. (2006b) should be approached with this in mind. Nevertheless, the consistencies of results presented below with those from Ungar et al. (2006b) give us confidence that the difference in sample composition has little effect on our interpretations.

As with the feature-based microwear analysis, the specimens used in this study come from Plio-Pleistocene deposits at Olduvai Gorge, Koobi Fora, Baringo, West Turkana, Sterkfontein and Swartkrans. These can be divided into four groups: Homo habilis (n = 5), Homo erectus (n = 8), Homo sp. indet. from Sterkfontein Member 5C (n = 2), and Homo sp. indet. from Swartkrans Member 1 (n = 3). The taxonomic attributions of the individual specimens are presented in Table 11.3. Explanations of and rationale for these assignments are presented by Ungar et al. (2006b) for specimens used in that study, and by Wood and Van Noten (1986) and Wood (1991) for the others.

Methods of specimen preparation followed conventional microwear procedures, except that replicas were not coated or mounted on stubs (as would be necessary for typical SEM study). Original fossils were cleaned with cotton swabs soaked in alcohol or acetone. Molds of occlusal crowns were prepared using President's Jet regular body polyvinylsil-oxane dental impression material (Coltene-Whaledent Corp) and casts were poured using Epotek 301 (Epoxy Technologies, Inc.) epoxy resin and hardener. Resulting epoxy replicas were mounted directly on the confocal microscope stage with plasticine.

Table 11.2 Descriptive microwear texture statistics

Taxon

Statistic

n

Complexity

Scale of maximum complexity

Anisotropy

Heterogeneity (=HAsfS cells)3

Textural fill volumeb

Alouatta

Mean

11

0.3603

53.4269

0.0058

0.6924

3871

palliata

Median

0.3149

0.2668

0.0057

0.5211

482

Standard deviation

0.1834

175.2870

0.0021

0.3827

4,583

Skewness

0.4020

3.3165

-0.3064

1.2619

0.7520

Cebus apellac

Mean

13

5.4658

31.9023

0.0037

0.7863

9,683

Median

2.8818

0.2666

0.0029

0.7603

9,707

Standard deviation

6.3043

91.1688

0.0019

0.3342

4,924

Skewness

1.6238

3.2418

0.7579

0.5457

0.7458

Lophocebus

Mean

15

1.7687

28.4170

0.0038

0.5350

12,369

albigena

Median

1.0181

7.6932

0.0035

0.5049

12,669

Standard deviation

1.7398

52.5858

0.0020

0.2764

2,374

Skewness

2.3044

2.7791

1.2128

1.6457

-0.4201

Trachypithecus

Mean

12

0.7337

1.2233

0.0048

0.6124

8,342

crístatus

Median

0.5141

0.2386

0.0036

0.5659

8,492

Standard deviation

0.6603

2.5286

0.0026

0.2855

4,365

Skewness

2.6182

3.0584

0.8262

0.8144

-0.4098

Early Homo

Mean

18

1.4173

1.0906

0.0034

0.4587

8,552

Median

1.0855

0.3451

0.0029

0.4228

10,096

Standard deviation

0.9365

2.0868

0.0019

0.1663

5,500

Skewness

1.3978

3.6264

0.5210

1.1522

-0.0363

Homo erectus

Mean

8

1.7339

0.4782

0.0037

0.4327

7,078

Median

1.4553

0.1808

0.0032

0.4144

6,220

Standard deviation

1.1968

0.7798

0.0015

0.1142

5,422

Skewness

0.8158

2.7827

0.3960

0.3386

0.8405

Homo habilis

Mean

5

0.9605

2.7615

0.0038

0.3717

13,006

Median

0.7469

1.3690

0.0042

0.3695

12,639

Standard deviation

0.4392

3.6889

0.0028

0.0696

2,994

Skewness

1.4656

1.9876

0.0476

-0.1320

0.7865

a Heterogeneity reported in Scott et al. (2006) was Hasfc9 some parts of the text erroneously referred to Hasfc81 . b Descriptive statistics for Tfv of extants species differ slightly from those reported in Scott et al. (2006) and were calculated and improved (non-directional) volume filling algorithm in Sfrax.

c The Cebus apella sample used here and in Scott et al. (2006) includes only specimens form Bahia, Brazil. to Cebus xanthostemos and Cebus nigritus robustus.

a Heterogeneity reported in Scott et al. (2006) was Hasfc9 some parts of the text erroneously referred to Hasfc81 . b Descriptive statistics for Tfv of extants species differ slightly from those reported in Scott et al. (2006) and were calculated and improved (non-directional) volume filling algorithm in Sfrax.

c The Cebus apella sample used here and in Scott et al. (2006) includes only specimens form Bahia, Brazil. to Cebus xanthostemos and Cebus nigritus robustus.

using the most recent These specimens have also been attributed

The microwear texture analysis followed procedures described by Scott et al. (2006). Three dimensional point clouds representing each specimen were generated using a Sensofar Pl|i confocal microscope (Solarius, Inc.) with an integrated white light vertical scanning interferometer (see Fig. 11.4). Data were collected for four adjacent fields on a "phase II" facet (usually facet 9) using a 100x long working distance objective. This generated about 1,738,000 elevations for each surface analyzed, sampled at intervals of 0.18 |im along both the x- and y-axes with vertical resolution specified to be five nanometers (0.005 |im). The combined work envelope of the fields examined was 276 x 204 |im. This resolution is better and work envelope is larger than those reported for the recent SEM-based microwear study of these hominins (e.g., Ungar et al., 2006b).

Resulting point clouds were analyzed using scale-sensitive fractal analysis (SSFA) software (ToothFrax and SFrax,

Surfract Corp). The premise of SSFA is that a given surface can look different at different scales - an asphalt road may look smooth at coarse scales (to a motorist driving along it), but rough and bumpy at finer scales (to an ant trying to cross it). Several SSFA texture attributes identified as useful for microwear analysis were considered for early Homo. These include area-scale fractal complexity (Asfc), anisotropy (epLsar18 |m), scale of maximal complexity (Smc), textural fill volume (Tfv), and heterogeneity of complexity (HAsfc). Each is described in detail elsewhere (Ungar et al., 2003, 2007; Scott et al., 2006).

Area-scale fractal complexity is change in roughness of a surface across scales of observation. The faster a measured surface area increases with scale, the higher the Asfc. Anisotropy can be measured as variation in lengths of transect lines sampled at given scales across surfaces in different orientations. A highly anisotropic scratched surface, for example, will have shorter transects when sampled parallel to

A. paIiiata ct T. cristatus ate a early Homo mbkx « x /.. albigena •»

A. palliata o o <d o o oq o T. cria latus aaaa/a a a a a early Homo äxxxx x x

0.0000 0.0040 0.0080

Anisotropy lepl.sar |(!mJ

0.0120

A. palliata CD O O OO O T. cristatus & a a a & a4 a early Homo ;-x x x /.. albigena C. apella x x xx xs£>x 2xx ♦ e<3as>s ♦ ♦

0 5000 10000 15000

Textural fill volume [Tfv]

20000

A. palliata T. cristatus early Homo L. albigena C. apella co cxm oo , aaafíi a

0.00 0.20 0.40 0.60 0.80 1.00 1.20 Heterogeneity |HAsfc5c[]J

Fig. 11.3 Distributions of the microwear texture values of (a) complexity, (b) anisotropy, (c) textural fill volume, and (d) heterogeneity for extant the comparative samples and the combined early Homo sample.

Table 11.3 Microwear texture data for individual early Homo specimens

Taxon

Specimen

Asfc

ePLsari^m

Smc

Tfv

HAsfCS cells

H. erectus

KNM-BK 8518

0.51

0.0035

0.35

1,552.14

0.57

H. erectus

KNM-ER 807

1.65

0.0023

0.21

10,049.87

0.46

H. erectus

KNM-ER 820

3.77

0.0055

0.15

17,395.20

0.33

H. erectus

KNM-WT 15000

1.26

0.0024

0.15

4,314.10

0.39

H. erectus

OH 60

2.06

0.0021

0.15

10,141.65

0.27

H. erectus

SK 15

0.62

0.0057

2.40

5,454.38

0.61

H. erectus

KNM-ER 1808

3.16

0.0052

0.15

6,985.56

0.42

H. erectus

KNM-ER 992

0.85

0.0030

0.27

729.81

0.41

H. habilis

OH 4

1.68

0.0042

2.34

12,639.34

0.44

H. habilis

OH 15

0.65

0.0054

0.27

17,438.06

0.37

H. habilis

OH 16

0.75

0.0011

0.62

10,307.38

0.29

H. habilis

OH 41

1.09

0.0008

1.37

14,302.75

0.32

H. habilis

Stw 19

0.65

0.0072

9.21

10,340.10

0.44

early Homo, Sterkfontein Mb. 5C

SE 1508

0.72

0.0060

0.43

11,922.00

0.35

early Homo, Sterkfontein Mb. 5C

SE 1579

1.04

0.0023

0.42

0.00

0.27

early Homo, Swartkrans Mb. 1

SK 27

0.84

0.0016

1.21

1,878.74

0.58

early Homo, Swartkrans Mb. 1

SK 45

1.48

0.0016

0.21

5,640.57

0.55

early Homo, Swartkrans Mb. 1

SK 847

1.28

0.0029

0.67

12,849.83

0.88

the preferred orientations of scratches than sampled across those scratches. Transects perpendicular to striations are longer because they must move in and out of individual features as they cut across them. Thus, heavily pitted surfaces tend toward high complexity and low anisotropy, whereas surfaces dominated by shallow, parallel striations are the opposite, with lower complexity and high anisotropy.

Other attributes allow us to fine-tune descriptions of surface texture. Scale of maximal complexity identifies the scale range over which Asfc is calculated, such that larger values for Smc should correspond to more, deeper features at coarser scales. Textural fill volume increases as more filling elements can be packed into features at a given scale. Textural fill volume tends to be larger as features become larger

Fig. 11.4 Three-dimensional photosimulations derived from microwear surface point clouds for (a) Alouatta palliata, (b) Cebus apella, (c) Trachypithecus cristatus, (d) Lophocebus albigena,

(e) Homo erectus, (f) Homo habilis, (g) early Homo Swartkrans Mb. 1, and (h) early Homo Sterkfontein Mb. 5C.

Fig. 11.4 Three-dimensional photosimulations derived from microwear surface point clouds for (a) Alouatta palliata, (b) Cebus apella, (c) Trachypithecus cristatus, (d) Lophocebus albigena,

(e) Homo erectus, (f) Homo habilis, (g) early Homo Swartkrans Mb. 1, and (h) early Homo Sterkfontein Mb. 5C.

and/or more square or circular as opposed to linear. Heterogeneity of complexity is simply a measure of how much Asfc tends to vary across a given surface at a given scale. Thus, surfaces with wear that is similar in degree and type (homogeneous) will have lower values for HAsfc.

Early Homo specimens were compared first to a baseline series of extant taxa published recently by Scott and coauthors (2006), and then to one another. The extant baseline series included two pair of contrasting species, Cebus apella

(n = 13) and Alouatta palliata (n = 11) and Lophocebus albigena (n = 15) and Trachypithecus cristatus (n = 12). The techniques used for data collection were the same as those described here for early Homo. Details about these specimens can be found in Scott et al. (2006). Suffice it to say that these are all specimens are all wild-shot individuals, and that each pair contained a species considered to be frugivorous with hard-object components (C. apella, L. albigena) and a species reported to be more foliviorous (A. palliata,

T. cristatus). While a primate species that specializes on underground storage organs would be a valuable addition to the extant baseline, such microwear data are not currently available. Extant species examined here nevertheless demonstrate how primates with diets dominated by tough foods differ in microwear textures from those that consume more hard, brittle ones. And it is these fracture properties, rather than food types themselves, that microwear differences should reflect. It would be useful to examine dental microwear in terrestrial species too, as microwear patterns may also be affected by differences in exogenous grit related to substrate preferences (Daegling and Grine, 1999).

First, a MANOVA was performed on ranked data for all variables (Asfc, epLsarlg , Smc, Hasfcg celb and Tfv) to compare a combined sample of early Homo to the extant baseline taxa. Individual ANOVAs and multiple comparisons tests were used to determine the sources of significant variation. Both Tukey's Honestly Significant Difference (HSD) and Fisher's Least Significant Difference (LSD) tests were used to balance risks of Type I and Type II errors (Cook and Farewell, 1996). Values of p < 0.05 for Tukey's HSD tests may be assigned significance with some confidence, whereas values of p < 0.05 on Fisher's LSD tests but not Tukey's HSD tests are here considered to suggest possible but unimpressive differences between pairs.

Early Homo specimens were then divided into four groups following Ungar et al. (2006b): (1) Homo habilis; (2) H. erectus; (3) Homo from Sterkfontein Mb. 5C; and (4) Homo from. Swartkrans Mb. 1. These groups were compared using the same scheme described above - a MANOVA on ranked data for all variables, and individual ANOVAs and multiple comparisons tests to determine sources of significant variation. While expectations of significant variation between the groups are admittedly optimistic given the small samples and resulting low power, significant variation in pit percentage was reported for the conventional microwear study (Ungar et al., 2006b), thus differences in microwear textures could indicate directions for future research.

Was this article helpful?

0 0

Post a comment