Analysis of bone geometry

As outlined in section 4.5.2 bone size and shape have the ability to inform about changes in social organisation and activity patterns. Accordingly, long bone

measurements were taken to explore whether the emergence of Islam impacted on the existing social structure through an analysis of changing bone size. This was achieved by looking at both long-bone lengths and cross-sectional indices (bone loading). While it would have been preferable to use imaging or x-ray, absolute measurements and indices still provide a useful mode of analysis as they have been shown to produce useful indicators of change.

Specifically, the analysis of absolute measurements was important for three reasons.

The analysis of sexual dimorphism and temporal dimorphism can highlight whether there were differences in activity between men and women, and between temporal comparison groups set out in section 6.3.1. It can also demonstrate general changes in social organisation. Understanding the sex difference and temporal variation in absolute measurements was also crucial for identifying where differences in body size may have impacted on entheseal results.

Indices acted as a measure of robusticity and identified whether loading on the clavicle, humerus, ulna and radius differed between groups or over time. Absolute measurements were analysed according to the comparison groups outlined in section 6.3.1, prior to

indices. Due to fragmentation in the Visigothic sample, indices were only analysed for the Islamic comparisons. Analysis of stature did not make up part of this research as this was already investigated by Pomeroy and Zakrzewski (2009).

6.4.4.1 Measurements data collection

Absolute measurements

Measurements were taken from specific fixed anatomical landmarks to ensure comparability between individuals. Standard measurements outlined in Buikstra and Ubelaker (1994) were followed. These measurements are based on the work by Moore-Jansen et al. (1994) using skeletal landmarks identified by Martin and Seller (1957).

The measurements are frequently used in osteoarchaeology, permitting comparison with other samples. Specific measurements used in this research are presented in Table 6.10 Left and right elements were measured where possible. Broken long bones were only measured if fragmented into 2 parts, if the break was clean, and could be held together securely and measured with the assistance of a second party.

Bone lengths were measured using an osteometric board. Diameters were taken using electronic sliding callipers. A tape measure was used to take circumferences. Means and standard deviations were calculated for each measurement. All measurements were taken in mm to the nearest 2.d.p. Measurements were taken twice to ensure accuracy.

A particularly good indicator of social change is fluctuations in the degree of sexual dimorphism. A number of mathematical methods exist for the calculation of sexual dimorphism, including ratios and regression analysis (see Smith 1999). While

regression analysis has been used in biological studies, it has been shown to be no more beneficial than ratios (Smith 1999:453). Accordingly, sexual dimorphism (DI) was calculated by dividing the male mean by the female mean, as proposed by Smith (1999), following Lovich and Gibbons (1992). The result indicates the male measurement as a proportion of the female mean.

Table 6.10. Definitions and sources for measurements used in this research

Measurement Abbreviation Source

Clavicle

Superior inferior diameter of sternal clavicle DsistC Derived Anterior posterior diameter of the sternal clavicle DapstC Derived Superior inferior diameter of the acromial end of the clavicle DsiacC Derived Anterior posterior diameter of the acromial end of the clavicle DapacC Derived

Maximum clavicle length XLC Buikstra and Ubelaker (1994)

Circumference of clavicle midshaft CmC Buikstra and Ubelaker (1994)

Superior inferior diameter at clavicle midshaft DsimC Buikstra and Ubelaker (1994) Anterior posterior diameter at clavicle midshaft DapmC Buikstra and Ubelaker (1994) Humerus

Maximum length of the humerus XLH Buikstra and Ubelaker (1994)

Physiological length of the humerus LphH Martin and Seller (1957) Minimum circumference of humeral shaft YCH Martin and Seller (1957)

Circumference of humeral head CHH Martin and Seller (1957)

Maximum diameter of the humeral midshaft XDmH Buikstra and Ubelaker (1994) Minimum diameter of the humeral midshaft YDmH Buikstra and Ubelaker (1994) Sagittal diameter of the humeral head XDsHH Martin and Seller (1957) Transverse diameter of humeral head XDtHH Buikstra and Ubelaker (1994) Width of the humeral proximal epiphysis WPH Martin and Seller (1957) Width of the humeral distal epiphysis WDH Buikstra and Ubelaker (1994) Radius

Maximum length of the radius XLR Buikstra and Ubelaker (1994)

Transverse diameter at radial midshaft DtmR Buikstra and Ubelaker (1994) Anterior posterior diameter at radial midshaft DapmR Buikstra and Ubelaker (1994)

Transverse diameter of radial head DthR Martin and Seller (1957)

Maximum diameter of radial head DXhR

Width of distal radial epiphysis XWdR Martin and Seller (1957)

Ulna

Maximum length of the radius XLU Buikstra and Ubelaker (1994)

Anterior posterior diameter at ulna midshaft DapmU Buikstra and Ubelaker (1994) Transverse diameter at ulna midshaft DtmU Buikstra and Ubelaker (1994)

Patterns of long bone asymmetry calculated from absolute measurements have been used to explore differences in activity between populations (see section 4.5.3.2). For example, urban populations have been demonstrated to have more asymmetry than rural populations (Mays 2010). The difference between left and right measurements was calculated in SPSS.

Indices calculated from absolute measurements

Table 6.11 Indices calculated from raw measurements

Index Description Abbreviation Source

Clavicle sternal end superior-inferior diameter of sternal end

anterior-posterior diameter of sternal end Sternal SIAP Derived Clavicle acromial end superior-inferior diameter of acromial end

anterior-posterior diameter of acromial end

Acromial

SIAP Derived

Clavicle midshaft superior-inferior diameter midshaft

anterior-posterior diameter of midshaft -

Pomeroy and Zakrzewski (2009) Humeral midshaft minimum diameter of humeral midshaft

maximum diameter of humeral midshaft - Bridges et al.

(2000) Radial midshaft minimum diameter of radial midshaft

maximum diameter of radial midshaft - Wanner et al.

(2007) Ulna midshaft minimum diameter of ulna midshaft

maximum diameter of ulna midshaft -

Pomeroy and Zakrzewski (2009)

Cross-sections were analysed in order to compare whether loading on upper limbs varied between groups. Cross-section data was obtained by creating indices from the absolute measurements. Indices were calculated using the formulae outlined in Table 6.11. The indices were selected as they reflect bending strength and therefore the degree of force imposed on the bone. As no indices have been created for the sternal and acromial ends of the clavicle, superior-inferior dimensions were divided by anterior posterior measurement as is practiced for other long bone indices.

6.4.4.2 Statistical analysis of measurement data

Prior to comparative analysis, it was important that data was tested for normality to ensure the integrity of any significant results identified. Where data was normally distributed, measurements and indices were compared using a GLA test. Where data log transformed, it was retested and the P value marked with a *. If the data was not

normally distributed, a Mann-Whitney U test (a non-parametric equivalent) was used to test for statistical significance and the resulting P value is marked ⁺.

The degree and significance of asymmetry between left and right sides was tested using a repeated measure test (RPM). Where data did not transform to a normal distribution, a Kruskall-Wallis test was used on the difference between left and right estimates and the

P value is marked ^. A comparison of the pattern of asymmetry between males and females was also calculated using a RPM test or Kruskall-Wallis test.