Methods for inferring paleohabitats from the functional morphology of bovid phalanges

The functional morphology of postcranial remains can be used to infer habitat preference. This approach is typically considered an “ecomorphological” method, and has frequently been applied to bovid postcranial remains. Methods for predicting habitat preference from bovid postcrania currently exist for femora, metapodials, and astragali. Here we describe three methods for predicting paleohabitats using measurements of African bovid phalanges (proximal, intermediate, and distal). The proximal phalanx method correctly predicted the habitat preference for 130 of 183 modern bovid specimens (71.0%, 2.8 times better than chance, p < 0.0001). The intermediate phalanx method correctly predicted the habitat preference for 115 of 163 modern bovid specimens (70.6%, 2.8 times better than chance, p < 0.0001). The distal phalanx method correctly predicted the habitat preference for 87 of 122 modern bovid specimens (71.3%, 2.8 times better than chance, p < 0.0001). These accuracies compare well with those of existing such methods (1.8–3.4 times better than chance). Analysis of the probabilities associated with the habitat predictions allows confidence thresholds to be established that identify specific predictions which have <5% chance of being in error. This raises the effective accuracy of the methods to 95%. Extensive exploration and manipulation of the underlying data demonstrate that the habitat predictions are generally robust, and are relatively independent of body weight, taxonomy, and sample composition. These methods are broadly applicable, relatively accurate, and can be used to generate independent predictions of habitat from different elements, and thus constitute a useful approach to inferring past environments.     

An exploration of character traits and linear measurements for sexing mountain gazelle (Gazella gazella) skeletons

The skeletal elements of mountain gazelle (Gazella gazella) have proven difficult to sex with statistical confidence due to wide overlap in the body size of the two sexes. We studied a sample of 53 modern mountain gazelle skeletons to determine which character traits and metric measurements best predict sex. The success of the character traits was determined using blind tests while the metrics were examined using discriminant function analysis. The most useful elements include the previously identified horn core, pubis and atlas, but also some new bone portions that preserve well in the archaeological record (e.g., distal tibia, distal metacarpal and metatarsal, and second phalanx). Surprisingly, two elements commonly used in sexing analyses (distal humerus and astragalus) were not among the most effective elements. Although cutting points and discriminant functions for sexing gazelle bone portions are presented here, they do not account for potential body size change and thus are not suited for direct application to archaeological assemblages. Instead, we provide guidelines for application to archaeological gazelle assemblages, most importantly a regression analysis that considers the sex ratios obtained from multiple measurements to predict the sex ratio of archaeological gazelle populations.     

Methods for inferring paleohabitats from discrete traits of the bovid postcranial skeleton

Various “ecomorphological” methods exist for using the functional morphology of bovid postcranial remains to reconstruct paleohabitats. Most such methods use measurements, but both Gentry [The Bovidae (Mammalia) of the Fort Ternan fossil fauna, in: L.S.B. Leakey, R.J.G. Savage (Eds.), Fossil Vertebrates of Africa, vol. 2, Academic Press, London, 1970, pp. 243–323] and Köhler [Skeleton and Habitat of Recent and Fossil Ruminants, Münchner Geowissenschaftliche Abdhandlungen 25 (1993) 1–88] have identified numerous discrete (non-metric) traits of the bovid postcranial skeleton that are said to be indicative of habitat preference. However, these traits have not been systematically tested on a modern bovid sample. We report here such a test. Eighty-six non-metric characters were evaluated using a sample of modern African bovids (n = 197). Of the 86 characters, 48 were either insufficiently defined or exhibited too much intra-individual variation to be potentially indicative of habitat. Two characters were invariant in the sample. Of the remaining 36 characters, 11 were sufficiently correlated with habitat preference (Cramer's V > 0.5) to be of some potential use in reconstructing paleohabitats. These characters are primarily concentrated on the phalanges, and provide a means by which fragmentary phalanges can be used for habitat reconstruction, albeit at a broad level of resolution (open or closed habitat). The estimated accuracy of these methods is greater than 80%. Their use on fragmentary remains may lessen the bias introduced by basing habitat reconstructions on more complete fossils.     

Baboon Bone Mineral Densities: Implications for the Taphonomy of Primate Skeletons in South African Cave Sites

C. K. Brain (e.g., 1981) documented an interesting difference in relative skeletal part representation between primates and bovids of similar live body size recovered from the fossil cave site of Swartkrans (South Africa). Hominids and baboons are represented primarily by skull parts and a paucity of postcranial bones, while small sized bovids (Size Classes 1 & 2) are represented by more equivalent abundances of all skeletal elements. Brain argued that the Swartkrans bones were primarily carnivore collected, and postulated that the difference in element frequencies between primates and bovids is the manifestation of less durable primate postcranial skeletons relative to those of bovids when either are subjected to a destructive process such as carnivore feeding. Experiments by Brain (1981) in which baboon and bovid carcasses were fed to large, African carnivores lend support to this hypothesis. Bovid postcranial elements survived carnivore feeding more frequently and more completely than those of primates. Assuming that bulk bone mineral density (bulk BMD) is a measure of durability, we supplement Brain's observational data by presenting the first systematic, element-by-element comparison of baboon and bovid postcranial bulk BMD measurements obtained by dual energy X-ray absorptiometry (DXA). While the bovid in our sample exhibits greater absolute bulk BMD values in 57·8% of all measured bone areas, when grouping areas by body region, a statistically significant difference in absolute bulk BMD values between the baboons and bovids is documented only in the hindlimb region (i.e., femur, tibia and patella). Density differences in other body regions are not statistically significant. Further, in only one case (i.e., small bovids in the Member 2 [1948–1953/1965–1975 excavation] assemblage) is there a significant and positive correlation between skeletal part frequencies of primates or bovids and bulk BMD. This suggests that, for the most part, primate and bovid skeletal part representation at Swartkrans is not the result of density-mediated processes. Differences in skeletal part representation between primates and bovids at Swartkrans may thus be attributable to factors other than bone density—such as bone size, length, shape, and/or the relative palatability of surrounding soft tissues on bones.     

The Nature and Function of Astragalus Bones from Archaeological Contexts in the Levant and Eastern Mediterranean

Worked and unworked astragalus bones from a variety of different animals have been excavated throughout the eastern Mediterranean and Near East in several different contexts. This paper examines the nature and function of astragali and their modifications, and concludes that they were most often used as gamepieces, but also served a ritual function in cultic and funerary contexts. The frequent use of astragali in ritual contexts is due to the manner in which animals were skinned, and as a result of cultural diffusion. The frequent occurrence of worked astragali in cultic and funerary contexts allows them to serve as potential indicators of ritual activity in archaeological excavations.     

Discriminant function analyses in archaeology: are classification rates too good to be true?

The use of discriminant function analyses (DFA) in archaeological and related research is on the increase, however many of the assumptions of this method receive a mixed treatment in the literature. Statisticians frequently use complex statistical models to investigate analytical parameters, but such idealised datasets may be hard to relate to “real-life” examples and the literature difficult to assess. Using two faunal datasets that are more typical of archaeological and related research, one comprised of size-corrected linear measurements of bovid humeri and another of 3D geometric morphometric (GMM) shape data of African monkey skulls, and two simulated datasets, we illustrate some of the most important but often ignored issues of DFA. We specifically show why it is paramount to address “over-fitting” by cross-validation when applying this method and how the probability of correctly classifying cases by chance can be properly and explicitly taken into account.     

Morphometric identification of bovid metapodials to genus and implications for taxon-free habitat reconstruction

Principal Components Analysis (PCA) of measurements on twelve metacarpal and ten metatarsal dimensions indicates that metapodials differ much more in size than in shape among the twenty-three genera of extant Sub-Saharan bovids. The only conspicuous shape variation is in metapodial length, which sometimes differs significantly among genera that are similar in all other dimensions. It follows that Discriminant Analysis (DA) intended to identify metapodials to genus will depend primarily on size, and except for the occasional like-sized genera that differ in metapodial length, DA will often confuse genera of similar size. DA of the same metapodials subjected to PCA supports this prediction. PCA further shows that bovid species that favor the same habitat often differ in metapodial size, shape, or both, and species that favor different habitats often overlap substantially in their metapodial morphometrics. DA intended to assign metapodials to habitat will thus produce different results depending on the species used to characterize each habitat type and on the number of specimens by which each species is represented. Habitat reconstruction founded in DA is therefore taxon dependent rather than taxon-free, and it cannot supplant reconstruction based on taxonomic identifications below the family level.     

THE ANALYSIS OF SEA TURTLE AND BOVID KERATIN ARTEFACTS USING DRIFT SPECTROSCOPY AND DISCRIMINANT ANALYSIS*

We investigated the utility of diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) for the analysis and identification of sea turtle (Family Cheloniidae) and bovid (Family Bovidae) keratins, commonly used to manufacture historic artefacts. Spectral libraries are helpful in determining the class of the material (i.e., keratin versus plastics), but do not allow for inferences about the species source of keratin. Mathematical post‐processing of the spectra employing discriminant analysis provided a useful statistical tool to differentiate tortoiseshell from bovid horn keratin. All keratin standards used in this study (n = 35 Bovidae; n = 24 Cheloniidae) were correctly classified with the discriminant analysis. A resulting performance index of 95.7% shows that DRIFTS, combined with discriminant analysis, is a powerful quantitative technique for distinguishing sea turtle and bovid keratins commonly encountered in museum collections and the modern wildlife trade.     

A study of parametric versus non-parametric methods for predicting paleohabitat from Southeast Asian Bovid astragali

Studies of the African members of the family Bovidae have demonstrated the feasibility of using linear discriminant analysis to predict paleoenvironment based on measurements from postcranial skeletal elements. The use of discriminant analysis depends upon adherence to certain assumptions about the data set. In certain cases, these assumptions are violated, thus reducing the confidence of the results. The research presented here introduces the use of the non-parametric recursive partitioning method as a test of the linear discriminant method for predicting habitat from a collection of modern Bovid astragali from Southeast Asia. This study also provides a critique of the previous African studies. The results indicate that a combination of the parametric and non-parametric methods provides the highest prediction accuracy for Southeast Asian Bovids, though the independent methods do test favorably; thus providing an additional tool for archaeologists and paleoanthropologists in this region.     

FOCUS: The Scapula Representation could be the Key: A Further Contribution to the ‘Klasies Pattern’ Debate

Past discussion on the unusual skeletal part representations at Klasies River Mouth is briefly summarized. Recent discussion in this journal, regarding the “Klasies Pattern”, has focused upon the differential destruction of small and large bovid bone epiphyses by carnivore ravaging and density-mediated attrition. Bartram & Marean (1999) argue, from ethnoarchaeological study and consideration of other archaeological sites, that, unless shaft fragments are painstakingly identified, the upper limb bone epiphyses of large bovids will be seriously under-represented. They therefore suggest that the “Klasies Pattern” is likely to be artefact of taphonomic and analytical processes. Klein, Cruz-Uribe & Milo (1999) replied with a defence of the analytical procedures employed during the original Klasies River Mouth analysis. They also state that there was very little evidence of carnivore ravaging at Klasies River Mouth. In this paper, it is pointed out that Bartram & Marean's (1999) study only considered the humerus, radius, femur, tibia and metapodia. However, in the “Klasies Pattern” it is the scapula that is most notably abundant in the small bovid classes and most notably scarce in the large bovid classes. It is argued that, from the study of bone mineral densities and Brain's (1981) carnivore ravaging experiment, there is no reason to expect a differentially greater taphonomic destruction of large bovid scapulae. In fact, exactly the reverse may be true. It is therefore argued that at least this aspect of the “Klasies Pattern” must be considered to represent human differential bone transport, rather than an artefact of taphonomic processes.     

Sex estimation using cervical dental measurements in an archaeological population from Iran

Sex estimation of skeletal remains is one of the major components of forensic identification of unknown individuals. Teeth are a potential source of information on sex and are often recovered in archaeological or forensic contexts due to their post-mortem longevity. Currently, there is limited data on dental sexual dimorphism of archaeological populations from Iran. This paper represents the first study to provide a dental sex estimation method for Iron Age populations. The current study was conducted on the skeletal remains of 143 adults from two Iron Age populations in close temporal and geographic proximity in the Solduz Valley (West Azerbaijan Province of Iran). Mesiodistal and buccolingual cervical measurements of 1334 maxillary and mandibular teeth were used to investigate the degree of sexual dimorphism in permanent dentition and to assess their applicability in sex estimation. Data was analysed using discriminant function analysis (SPSS 23), and posterior probabilities were calculated for all produced formulae. The results showed that incisors and canines were the most sexually dimorphic teeth, providing percentages of correct sex classification between 86.4 and 100 % depending on the measurement used. The combination of canines and other teeth improved significantly the level of correct sex classification. The highest percentages of sex classification were obtained by the combination of canines and incisors (100 %) and canines and molars (92.3 %). The present study provided the first reference standards for sex estimation using odontometric data in an Iranian archaeological population. Cervical measurements were found to be of value for sex assessment, and the method presented here can be a useful tool for establishing accurate demographic data from skeletal remains of the Iron Age from Iran.     

Universal methodology for developing univariate sample-specific sex determination methods: an example using the epicondylar breadth of the humerus

In this paper we present a new methodology for developing sample-specific metric sex determination methods using unidentified individuals that can be applied to archaeological samples. This methodology does not require the assessment of sex of a large sample from pelvic morphological features as a first step in developing the method and instead is developed from a sample of individuals with sex unknown. The procedure involves using the overall mean of a measurement collected from the sample of unknowns as the discriminant criteria for determining the sex of the individuals in that same sample. Experiments with various sample sizes and sex ratios using distal humerus measurements from the Coimbra Collection (Portugal) suggest that allocation accuracies of 83–96% can be expected when the sample used to develop the method is greater than 40 individuals and the sex ratio is less than 1.5:1. The utility of the methodology is explored using the Lisbon Collection (Portugal) and an archaeological sample from Belleville (Canada) is tested as an example. The Lisbon Collection data indicate that joint measurements of long bones will provide the highest allocation accuracies. The results from the Belleville sample confirm that humerus joint measurements consistently provide the highest allocation accuracies (88–100%) even when the sex ratio in the sample used to develop the method is estimated at 1.9:1.     

Ascertaining an agent: using tooth pit data to determine the carnivore/s responsible for predation in cases of suspected big cat kills in an upland area of Britain

The identification of the involvement of a particular carnivore in the modification of bone assemblages concerns a number of fields of research including archaeological and palaeontological enquiry. Taphonomy provides a methodology by which bone assemblages can be analysed and interpreted and this is more often undertaken with archaeological or palaeontological assemblages. A taphonomic analysis is undertaken here in order to determine the perpetrator of predation attacks on domestic stock from a modern-day setting. Recently reported techniques using tooth marks preserved on bone surfaces made by known carnivores are successful at determining some class sizes of predators and are used here to determine the perpetrator(s). Although a class size of carnivore is readily identified by this methodology, a particular carnivore taxon is not. Tooth morphology and dental configuration are reported here as better criteria for identifying a particular taphonomic agent. Tooth pit dimensions are used here to identify the class size of carnivores involved, and tooth morphology and cusp spacing to suggest a medium sized felid and fox as taphonomic agents. The identification of the medium-sized felid may support observations and reports of alleged “big” cat kills in the area. The study has important implications for the interpretation of fossil sites where felids may have been involved in the modification of animal carcasses but are archaeologically invisible in terms of their fossil remains.     

Identifying artificially deformed crania

In this paper we report on a new discriminant function for the identification of artificially deformed crania. Development of the function, based on a sample of deformed and undeformed crania from the Philippines, required visual classification of the sample into deformed and undeformed groups. Working from the observation that deformed crania display flattened frontal and occipital regions, the sample was seriated based on degree of flattening; classification was based on the results of this seriation. The discriminant function, calculated using curvature indices, required only six simple measurements: arc and chord measurements for the frontal (glabella to bregma), parietals (bregma to lambda) and occipital (lambda to opisthion). The function was designed to be conservative, in that a deformed cranium may be classified as undeformed, but the opposite should not occur. Our function classified the undeformed crania with 100% accuracy and deformed crania with 76.9% accuracy, for a total of 91.9% agreement with visual classification. In order to evaluate whether the function is applicable for samples from outside the Philippines, a double blind test was conducted with a large sample of deformed and undeformed crania from a broad geographical and temporal range. For this sample, the function agreed with visual classification in 89.7% of cases; 98.8% of undeformed crania were correctly classified, while deformed crania were identified with 73.7% accuracy. These results demonstrate the utility of the new discriminant function for the classification of artificially deformed crania from diverse contexts. Copyright © 2007 John Wiley & Sons, Ltd.     

Geometric morphometrics and the population diversity of Late Glacial horses in Western Europe (Equus caballus arcelini): phylogeographic and anthropological implications

In order to test the possibility of large migrations of wild horses between the North and the South of Western Europe during the Late Glacial, we addressed the population diversity in this area, especially during the Bölling/Alleröd periods. We based this appreciation upon the detailed morphology of the distal part of the metapodials. Landmark approaches were applied to archaeological and palaeontological samples from three distinct areas: Switzerland Plateau, Paris Basin, and Charente (France). Little differences of isometric size characterized the three regional samples. Multivariate analyses of shape (principal component analysis, discriminant, canonical variates and neural network analyses) revealed similar sets of complex shape features on both metacarpals and metatarsals. In addition, regional groups appeared clearly differentiated by well defined shape patterns, the functional implications of which remain to be established. These results evidence the existence of a regional structuration of populations (particularly clear on the CVA of both metacarpals and metatarsals) suggesting the absence of long distance migrations. Finally, it appears that the distinction between size and shape systematically operated by geometric morphometrics can provide better insights into the study of the mobility of ancient populations. Geometric morphometrics approaches such as 3D Procrustes superimpositions, appear therefore to be of great interest for archaeological purposes.     

Actualistic research into dynamic impact and its implications for understanding differential bone fragmentation and survivorship

The relationship between bone mineral density and archaeological bone survivorship has played a critical role in zooarchaeological and taphonomic studies in recent decades. Numerous studies have suggested that higher-density skeletal element portions survive more frequently than lower-density element portions when archaeological assemblages are affected by some taphonomic processes. Interpretations of density mediated destruction have become commonplace in the archaeological literature, and are often used to explain the absence of certain bone elements and element parts in zooarchaeological assemblages. This study explores the effects of rockfall on bovid elements in varied environmental conditions and the differential survivorship of their element parts, and has implications for understanding the taphonomic processes through which bones are subjected to dynamic loading. Actualistic rockfall experiments conducted on twelve samples of frozen, fresh, and semi-dried bovid bones reveal that the generally low-density epiphyseal ends of bone elements resist fracture and analytical deletion with more frequency than the higher-density diaphyses. This evidence suggests that bone density does not correlate with likelihood of breakage or effective archaeological “destruction” when rockfall and other processes that result in dynamic impact are in action. While this research does not question the relationship between bone mineral density and the likelihood for archaeological survivorship as the result of some taphonomic processes, it presents one specific set of taphonomic processes that result in the differential survivorship of low density bone elements parts and the fragmentation and destruction of higher density element parts. This research presents evidence that shows that dynamic impact is a process capable of fragmenting and sometimes destroying high-density elements while low-density elements survive.     

Identifying bilateral pairs of deer (Odocoileus sp.) bones: how symmetrical is symmetrical enough?

Identifying bilaterally paired bones in zooarchaeological collections rests on the assumption that left and right skeletal elements from the same organism will be symmetrical. Bilaterally paired bones in an individual are, however, typically asymmetrical to some degree, demanding that the question “How symmetrical is symmetrical enough to identify a bilateral pair?” be answered with control data. The degree of symmetry chosen, or tolerance, will influence both how many true pairs in an archaeological collection are not identified (type I error) and how many false pairs are identified (type II error). Bivariate measures of 60 pairs of astragali and 48 pairs of distal humeri of deer (Odocoileus virginianus and O. hemionus) indicate that both sorts of error are frequent even when a conservative level of tolerance (≤average asymmetry) is used. Simulation of an archaeological collection indicates that as sample size increases, frequencies of both kinds of error increase. Application of the matching criteria and tolerance level to an archaeological collection underscores that the analytical requirements of identifying paired bones are steep.     

The effects of sample size on some derived measures in vertebrate faunal analysis

Analysis of three kinds of measures derived from archaeological and palaeontological faunal collections (measures of relative taxonomic abundance, measures of taxonomic diversity, and a measure proposed to distinguish natural from cultural bone) demonstrates that in many instances such measures may be, or are, a function of sample size. A procedure for detecting this situation is suggested, and the source of the interrelationships discussed.     

Occipital perforations in a late neolithic probable aurochs (Bos primigenius bojanus) cranium from Letchworth,Hertfordshire, UK

An example of ante‐mortem occipital perforations in the cranium of a probable aurochs (Bos primigenius) cow from a late Neolithic archaeological site at Letchworth, Hertfordshire is presented. This is the second reported occurrence thus far of cranial perforations in a wild bovid and lends support to a congenital cause of the condition in archaeological domestic cattle skulls. Copyright © 2002 John Wiley & Sons, Ltd.