Archive | Geographic Variation RSS feed for this section

Databse Fun: Of Databases, Statistics and Isotopes

25 Nov

I know what you are thinking – what sort of misspelled title is that for a blog post? The answer is below dear reader (1)!

Databases are, in my humble opinion, awful, tedious and time consuming beasts to create and are often best tackled head on armed only with a black coffee for sustenance as you try to accurately type a mind-numbing amount of data into an excel spreadsheet at 2am in the university library.  (That may just be my experience though!).  The beauty of a completed database, however, cannot be overestimated.  This is where you get to test out hypotheses based on the data that you have selected and gathered for your research question, where all of the core information lies and where the data can be repeatedly and demonstratively tested again and again.  A completed and ordered database is a thing of beauty and, when looked at 6am in the morning after a tiring night of inputting data, a thing of magnificence!

But let’s start at the beginning.  I recently had cause to look again at the database I had made for my MSc dissertation and, as I scrolled across and down the excel spreadsheet, I could just about remember the hours I had spent producing the spreadsheet, justifying the column titles and entering the data itself.  My data set included strontium isotopic results gathered from 422 individuals across 9 different sites from the Neolithic Linearbandkeramik (LBK, roughly 5500BC to 4800BC) culture of Central Europe, with my sample ranging geographically from the modern countries of Austria, Czech Republic and Germany. The data set used for my study was carefully culled from a literature review and a close reading of a number of journal articles that were available at that time (mid 2012).

My aim was to investigate statistically the claim of patrilocality in the LBK culture as proposed by Bentley et al. (2012) by investigating the specific sex and age differences within the profile group by using strontium isotopes as proxies.  Strontium isotopes samples (specifically 87Sr/86Sr) are often taken from both human and animal skeletal remains (primarily from teeth, specifically the 1st, 2nd and 3rd molars as they reflect Sr values throughout the life of an individual) as it survives well in archaeological contexts and is an informative approach to investigate mobility and local/non-local status of individuals.  Strontium values reflect geochemical signatures in the dietary component of the individuals, which comes from the soils and the underlying geological landscape that the individual lived on.  There are issues with this method (2) (see also this blog’s comments section).  Strontium isotopic investigations in archaeology are often studied in conjunction with oxygen isotopes (18O/16O) sampled from tooth enamel as well (specifically the 2nd molar) which represents water drank in life, but, frustratingly, this has not been the case in the LBK literature.

I knew that I wanted to statistically test the data set using SPSS 19, the standard statistical program widely used in the social sciences, but I first needed to tabulate and code the data so it would be useful when it came to testing the data.  As the study also included comparisons of the funerary grave goods and a basic demographic investigation of each site coding the entries (1=male, 2=female or 1=present 2=absent) allowed for comparisons to be made in the SPSS program and for statistical tests to be carried out.  The strontium itself was, as expected, non-parametric, which meant that the data adhered to no specific characteristic structure or parameter.


The normality test, using the Kolmogorov-Smirnov and Shapiro-Wilk statistical tests, indicates that the strontium data used for this study of the 422 individuals was not distributed normally (the P-value, nominally a significance value, is 0.000 for these tests). This means that tests such Spearman’s Rho correlation (quantity between variation), Mann-Whitney U (2 independent variables) and Kruskal-Wallis (3 or more independent variables) are the most appropriate statistical tests to perform on this data set (Bryman & Cramer 2011: 245).

When building the database I also wanted any relevant information and references easy to hand so I included the skeletal number (as given in the articles), site name, period, sex, sex code, isotope source, body position, funerary artefacts found and reference etc for each individual used in the study (see below).


A screen shot of the database used in my MSc dissertation displaying the revelent information of the 422 individuals from the LBK sites used in the study. The data was entered in a Excel spreadsheet before being transferred to SPSS for statistical investigation. Click to enlarge.

The data was carefully added over a number of days once I had gathered all the required journal articles discussing the sites I had chosen  The sites themselves were largely located in southern Germany, with the 9 sites nicely split into three time periods throughout the chronology of the LBK period.  Perhaps somewhat hastily I added this to the database and assigned the values of the individuals with a Early, Middle and Late ranking for their respective site.


Towards the bottom of the database used for the study. Here we can see the references cited for each site used in the study and the specific coding for funerary items (the two columns before reference column on the right hand side, where 1 depicts present and 0 absent).

During the construction of this database I did encounter problems as I had not built such a large database before, indeed the only time I had really used a database properly was for my undergraduate dissertation some years previous whilst using ArcGIS.  The problems this time included whether I was actually coding the funerary items the right way round or not, reading back through the database and correcting any errors in typing (especially for the strontium values) and making sure I correctly identifying the individuals used in their respective articles.  There are some things inherent in archaeology that cannot be solved.  This includes lacking contextual data or written site reports (which may or may not exist hidden in regional archaeological unit headquarters, not known or available to the public or indexed on any site).

Of course there were problems with my approach, which I expounded on in fuller detail in the thesis itself.  This did include problems interpreting the strontium results and distinguishing between local and non-local individuals at the site when there is no reference data to compare it to and debating my own statistical approach.  Still, as frustrating as building the database was, I did enjoy carrying out my own investigation of it immensely.  On rainy days I often think that my dataset could do with a second look at and investigation, perhaps I could change this approach or that, use this statistical method instead and isolate that clump of individuals etc.

It may be a pipe dream for the moment (I lack a working SPSS program for one!) but this is as much of a key part of archaeology and archaeological research as digging in the mud is.  Research is what drives archaeology and human osteology forward, from new scientific techniques to reviewing old data and finding new patterns.  The past is always present in new technology, you just have to drive it forward sometimes.

I will be introducing the Neolithic LBK culture in further detail in an upcoming post and discussing the merits of my thesis in further detail in another post.  For now I hope you have enjoyed this brief delve into what was the core of that research, the database itself.


(1.) This post was named in honour of a spelling mistake I made in the contents pages of my MSc thesis, spotted only when I proudly showed a friend a copy of the thesis a few weeks after the hand in date.  This, of course, led to gales of laughter from both of us (and to my internal cringing) as my poor editing skills came to light and it still remains a favoured joke to this day.

(2.) A few problems have become apparent with the strontium isotope technique, as with any mature and widespread application of a scientific technique, and it is worth mentioning them here (Bentley et al. 2004: 366).

Firstly is the issue of what a local and non-local signature mean for the prehistoric individual, as technically the 87Sr/86Sr ratio reflects diet over a period of time, and said food could have come from non-local sources.  However, this could be a distinct benefit, as it may be possible to identify individuals whose subsistence activity took place over a diverse range of territories (Bentley et al. 2004: 366, Price et al. 2002: 131).  Secondly, diagenesis affects anything buried and groundwater strontium has a tendency to penetrate the skeleton after burial (Bentley et al. 2004: 366).  In this study only enamel from the permanent dentition (1st or 2nd molars) is used, as this mitigates the effects of diagenesis because enamel is a strong biological material containing large mineral crystals, rendering it much less porous than bone and it is highly resistant to biochemical alteration (Killgrove 2010, Richards et al. 2008).  The third issue concerns the environmental heterogeneity of the strontium isotope signatures, which as Bentley (et al 2004: 366) points out ‘vary in different minerals of a single rock, in the leaves, stems and roots of a plant, or in water sources such as streams and precipitation’.  The measurement of small herbivore bones, or snail shells, at the locality of the archaeological site, preferably from the same chronological age, can obtain a remarkably consistent 87Sr/86Sr ratio, which is representative of the local catchment area (Bentley et al. 2004: 366).  The use of strontium ratio is however just one tool among many that is used to shed light on our ancestors; it should always be used in combination with other techniques of investigation to elucidate the full range of potential data present of archaeological sites and materials (Montgomery 2010, Richards et al. 2001, Van Klinken et al. 2000).


Bentley, R. A., Price, T. D. & Stephan, E. 2004. Determining the ‘local’ 87Sr/88Sr Range for Archaeological Skeletons: A Case Study from Neolithic Europe. Journal of Archaeological Science. 32 (4): 365-375.

Bentley, R. A., Bickle, P., Fibiger, L., Nowell, G. M., Dale C. W., Hedges, R. E. M., Hamiliton,. J., Wahl, J., Francken, M., Grupe, G., Lenneis, E., Teschler-Nicola, M., Arbogast, R-M., Hofmann, D. & Whittle, A. 2012. Community Differentiation and Kinship Among Europe’s First Farmers. Proceedings of the National Academy of Sciences Early Edition. doi:10.1073/pnas.1113710109. 1-5.

Bryman, A. & Cramer, D. 2011. Quantitative Data Analysis with IBM SPSS 17, 18 & 19: A Guide for Social Scientists. London: Psychology Press.

Killgrove, K. 2010. Migration and Mobility in Imperial Rome. PhD Thesis. University of North Carolina. (Open Access).

Montgomery, J. 2010. Passports from the Past: Investigating Human Dispersals Using Strontium Isotope Analysis of Tooth Enamel. Annals of Human Biology. 37: 325–346. (Open Access).

Price, T. D., Burton, J. H. & Bentley, R. A. 2002. The Characterisation of Biologically Available Strontium Isotope Ratios for the Study of Prehistoric Migration. Archaeometry. 44 (1): 117-135.

Richards, M.P., Fuller, B,. T. & Hedges, R. E. M. 2001. Sulphur Isotopic Variation in Ancient Bone Collagen from Europe: Implications for Human Palaeodiet, Residence Mobility, Modern Pollutant Studies. Earth and Planetary Science Letters. 191 (3-4): 185-190.

Richards, M. P., Montgomery, J., Nehlich, O. & Grimes, V. 2008. Isotopic Analysis of Humans and Animals from Vedrovice. Anthropologie. XLVI (2-3): 185-194.

Van Klinken, G., Richards, M. and Hedges, R. 2000. An Overview of Causes for Stable Isotopic Variations in Past European Human Populations: Environmental, Ecophysiological, and Cultural Effects. In S. Ambrose and M. Katzenberg (eds). Biogeochemical Approaches to Palaeodietary Analysis. New York: Kluwer Academic. pp. 39-63.

Skeletal Series Part 1: Bone Variation & Biomechanics

10 Apr

In the following series of blog posts I aim to cover each of the main skeletal elements.  Each post will have a single focus on a bony element, from the skull down to the bones of the foot.  Firstly though we must deal with the variation that human osteologists and bioarchaeologists will see in individual skeletons, and in a population series.  It is both useful and informative to learn the differences and the effects caused by the 4 main variation factors in the morphology of human bones.  As it is by ascertaining the degree of influences that the variation factors can cause that we can begin to understand the individual, and the skeletal series of a population, in a more informative and considered way.  The second part of this entry will focus on the basics of biomechanics, and the influences certain lifeways can have on bone.

The basic biology of bones was previously discussed in this post here, and of teeth here.  Bone in its natural state must be recognised as a changing living organism throughout life that responds to stress, both nutritional and mechanical, and remodels accordingly.  It is also must be remembered that bone is a composite material, and is able to heal itself.

Variation 1 : Ontogeny

Ontogeny is simply growth and development of an organism, in this case Homo sapiens.  The archaeological record of skeleton remains include unborn individuals right through to individuals in their 70th year and beyond.  Typically there are 7 classification groups of human age groups.  They run from Fetus (before birth), Infant (0-3 years), Child (3-12 years), Adolescent (12-20 years), Young Adult (20-35 years), Middle Adult (35-50 years), and Old Adult (50+ years) (White & Folkens 2005: 364).  Differences in bone structure, and in the growth of different bone elements often manifest themselves in changes in size and shape.

Adult with Two Juvenile Remains, From Southern Sahara

Basic Growth Profile for Homo Sapiens, Notice the Large Cranium and the Way the Body Catches Up.

Variation 2: Sexual Dimorphism

Humans are sexually dimorphic, that is there are differences between the female and male body size.  Although not as distinct between our cousins such as the gorillas, female human skeletal remains are relatively smaller in both bones and teeth size (Jurmain et a 2010).  Such skeletal variation is also manifest in the requirement of reproductive functions in the female skeleton, thus we are also often able to tell sex from skeletal remains (White & Folkens 2005: 32).

Generalised Male:Female Sexual Dimorphism

Variation 3: Idiosyncractic Differences

The idiosyncratic  (or individual) differences found in skeletons are simply natural variations, in the understanding that every body is different, and rarely are people exactly the same (identical twins excluded).  Idiosyncratic differences in bone affect the size and shape of the bone, and the topography of the bone surface.  Again, such variation is very common in human skeletal remains (White & Folkens 2005: 32).

Disarticulated human bone from the site of Armana, ancient Egypt.

Variation 4: Geographic or Population-Based

As White & Folkens point out ‘different human groups can differ in many skeletal and dental characteristics’ (2005: 32).  Thus this geographic variation can be employed to assess population affinities between skeletal series.  This trait can be quite useful in determining commingling of certain populations in prehistoric skeletal series as certain environmental and genetic traits can be passed on.

Biomechanic Basics

So these are the four main variations we should be aware of when we are looking and studying individual skeletons or a series of a population.  By considering these four main variations we can study the individual’s life pathway alongside other lines of investigation.  What we must also take into account next are the basics of biomechanics.  Biomechanics is the application of engineering principles to biological materials, whilst remembering that bone can remodel and change according to pressures put upon the bone.  As Larsen states that ‘the density of bone tissue differs within the skeleton and within individual bones in response to the varying mechanical demands’ (1997: 197).  It must be remembered that the response of human bone to ‘increased loading is in the distribution of bone (geometric) rather than density or any other intrinsic material property of bone’ (Larsen 1997: 197).

Importantly it is noted that Human bone is anisotropic, meaning its mechanical properties vary according to the direction of the load.  Importantly, Wolff’s Law highlights how bone replaces itself in the direction of functional demand.  A classic example of the remodelling capabilities of bone is that of the tennis player who has thicker cortical bone in their dominant arm.  This manifests itself in thicker cortical bone alongside hypertrophy of the muscle attachment sites.  One study carried out found that ‘males have a 35% increase in the cortical bone in the distal humerus of the playing arm vs the non-playing arm’, helping to exemplify Wolff’s Law (Larsen 1997: 196).  That study was an example of bilateral asymmetry humeral loading.  Alongside, it is the action of the main forces acting on human bone that help to change the bone, these  include a) compression, b) tension, c) shear, d) torsion & E) compression + tension+ bending.

Wolff’s law states that healthy load bearing bone (LBB) responds to strain by ‘placing or displacing themselves (at a mechanical level) in the direction of the functional pressure, & increase or decrease their mass to reflect the amount of functional pressure’, often muscular strain and/or weight bearing pressures (Mays 1999: 3).  As a part of this Frost (2004: 3) argues that the ‘mechanostat’, a tissue level negative feedback system, involves ‘two thresholds that make a bone’s strains determine its strength by switching on and off the biologic mechanisms that increase or decrease its strength’.  However, Skerry (2006: 123) has argued that there are many ‘mechanostats’ operating on the LBB and that different elements throughout the skeleton require different strain magnitudes for maintenance. Furthermore Skerry (2006: 126) also notes that differences are apparent between the sexes, and that genetic constitution, concomitant disease, exercise & activity patterns must be considered.

A recent article has also highlighted how the femoral neck width of obese people changes to accommodate the added weight.  In this case the width of the femoral neck has increased to dissipate weight throughout the bony area by increasing surface area and strength through redistribution of bone.  This is an example of active bone remodelling adapting to changes that the person has gone through in life.

An archaeological example of the above will now be taken from Larsen 1997.  ‘In the Pickwick Basin of northwestern Alabama, analysis carried out on both femora and humeri cross-sectional geometry has helped to reveal a number of differences between earlier Archaic Period hunter-gatherers and later Mississippian Period agriculturalists‘ (1997: 213).  From the femora measurements it seems that the both female and male agriculturalists had a greater bone strength, whilst analysis of male humeri shows little difference between the two series.  This has helped to show that activity levels increased for males but only in the lower limbs, as evidenced by the cross-section geometry.  However, for females of both time periods both humeri and femora strengths increased.  The article cited in Larsen (1997), Bridge 1991b, findings indicate that changes are from a greater range of activity undertaken by females than males.  With palaeopathological signs of osteoarthritis, it is concluded that the shift to food production,in particular maize production, may have had a relatively greater impact physically on women in this setting.

The next post will focus on ethics in human osteology, and from there we will consider each of the anatomical skeletal elements in context of their relative limb.


Frost, H. M. 2004. (A 2003). Update on Bone Physiology and Wolff’s law for Clinicians. Angle Orthodontist.  February 2004. 74 (1): 1-15.

Jurmain, R. Kilgore, L. & Trevathan, W.  2011. Essentials of Physical Anthropology International Edition. London: Wadworth.

Larsen, C. 1997. Bioarchaeology: Interpreting Behaviour From The Human Skeleton. Cambridge: Cambridge University Press.

Mays, S. 1999. The Archaeology of Human Bones.Glasgow: Bell & Bain Ltd.

Skerry, T. M. 2006. One Mechanostat or Many? Modifications of the Site-Specific Response of Bone to Mechanical Loading by Nature and Nurture. Journal of Musculoskeletal & Neuronal Interaction. 6 : 122-127. (Open Access).

White, T. & Folkens, P. 2005. The Human Bone Manual. London: Elsevier Academic Press.