Unlocking the Past

by Martin Jones

  To get closer to their identity, Gerstenberger started work on DNA sequences from their Y chromosome. She focused her attention upon four microsatellite regions along the chromosome. Nearest to the stones commemorating the first two earls, Hanns Cristoph and his son Hanns Sigismund, were two male burials with identical Y patterns, with tandem repeat lengths of 14, 23, 12 and 28 at the four regions studied. Moving to the next generation, two of Hanns Sigismund’s sons are commemorated in the chancel. One gave a poor amplification result, but the other, Wolf Ehrenreich, once again came up with 14, 23,12 and 28. The grave nearest to his son’s inscription was apparently robbed, but his grandson Josef Wilhelm followed, echoing the same genetic pattern. So far, so good, but this is where the bones began to deviate from the inscriptions.

  We know from contemporary records that Josef Wilhelm was very anxious to have a son and heir, and was rather relieved when Georg Josef turned up. He might, however, have been surprised by what Julia Gerstenberger would later discover. Georg Josef’s tandem repeat signature came out as 14, 24, 13 and 33. It is inconceivable that such a switch could have occurred in a single generation, or even in several generations. His true father was not the previous earl.

  This left the two female bodies, who obviously could not be studied through the Y chromosome. However, a study of autosomal microsatellites would bring them into the picture. Autosomes (the normal chromosomes) had the advantage of being found in both males and females, but, because they recombined with each generation, could not be used for the kind of lineage tracking possible with either mitochondrial or Y haplotypes. Their careful analysis would allow the women to be related to the male line, and Gerstenberger to trace this story to its conclusion.

  The development of the teeth and bones from the two female skeletons indicated different ages at death. One skeleton was of a mature woman of approximately thirty, the other of a girl of around thirteen years of age. What the autosomal microsatellites generated was a pattern consistent with the two females being the illegitimate earl’s wife and daughter. In other words, the microsatellites from the skeletons of the male and the older female could be recombined to fit the pattern found in the younger female. We know from historical records that this earl’s wife died at the age of thirty-six, matching the skeletal evidence. Both these women had died well before their time. The illegitimate earl had inadvertently broken the male line and now was consciously breaking with tradition. Having lost his wife when she was in her prime, and his daughter when she was still in her youth, he might not be able to give them a memorial inscription but he certainly wanted them laid within his ancestral mausoleum. The next earl chose to be buried elsewhere, and the male line in any case came to an end in 1815. Their ancestors had left behind two stories, the first a formal story inscribed in the chancel walls, the second a real-life story hidden within the DNA. The first story was of a traditional ideal, of a patrilineage of upright, aristocratic men. The second story was of real people, of yearning for an heir, of an extra-marital liaison, and of a funerary tradition defied in the throes of love and grief.

  when the names fall silent

  How much further could such narratives be taken back in time? In principle, they could go back as far as the recorded genealogies, which in every case is well within the longevity of the short strands of ancient DNA. If it is only the male lines that are documented in full, then more relevant than mitochondrial haplotypes are the Y haplotypes, and research into these is now growing fast. Is there anything to be gained by going back beyond the time scale of these genealogies and examining relationships of kin in prehistory?

  The problem might be compared with tree-ring dating and its calibration. The first tree-ring patterns to be fixed to a particular year were comfortably within the historical time scale. They were sequenced records from timber-framed buildings and wooden waterfronts whose dates of construction were securely documented, providing anchor points in the historical record for the patterns recorded in ancient wood. Back in prehistory, these historical anchor points are lacking. At first, prehistoric tree-ring patterns were assembled into floating chronologies, isolated local patterns without an absolute date. In time, these floating chronologies grew to overlap, first with each other, and then with the historically fixed chronologies, and the whole master chronology came together to provide prehistory’s most precise dating framework by far.

  This is only a loose analogy, and we will probably never achieve comparable master genealogies for prehistory. However, the floating chronology does provide a model of what might be achievable with current methods, a series of floating genealogies, with no actual names to anchor them, but with their interrelationships established. Such an approach might tell us a great deal about past societies through their burial sites. Who was placed in the ground near to whom? Were nuclear families, or larger groups, buried together, or were age and gender more relevant? In Gottingen, Susanne Hummel and Bernd Herrmann decided to look in the nameless archaeological record for something equivalent to the Romanov group. They found it in a cluster of bodies from a burial pit dug in the first millennium AD on the northern Rhine at Kleve, close to the German border with Holland. Excavation had exposed the skeletons of a mature man and woman, neatly laid alongside each other. Between and around them were the remains of the skeletons of three children. The fusion of the children’s bones suggested ages ranging between one and ten years. They looked the ideal test for a family burial, to be assayed by microsatellite studies of ancient DNA. Using microsatellites similar to those examined in the forensic cases discussed above, the team established that at least three of the five bodies were close relatives. In a similar approach to that taken in the case of the earls of Konigsfeld, they used the ancient DNA to establish both kin relations and the sex of the burials. Putting these together, the five bodies could be interpreted as a mother and father and their three children–a daughter of eight to ten years of age, a son of one or two years, and a third child, whose remains were the only ones too poorly preserved for analysis.

  Family structures are always interesting, but especially so when society is in the process of considerable upheaval. At such times the balance between the clan, the extended family, the nuclear family and so on may change as a major and integral part of the wider changes in society. Take, for example, the transition between foraging and farming that has been a recurrent theme in biomolecular archaeology. A mobile hunting group and settled village may organize their families and relations of kin in quite different ways, and understanding those changes is central to understanding the transition as a whole.

  Half a century ago, the anthropologist George Peter Murdock looked for a global logic in the relationship between kin patterns and the broader organization of society. Examining a large number of living and recent societies from around the world, he explored kinship relationships, the ways in which property was passed down the family line, and the norms that determined which relatives lived near to which. He went on to relate these patterns to other social and economic attributes of each community. One pattern that captured his attention was a trend for patterns of inheritance and residence to be linked. If goods passed down the male line, then more often than not the male line stayed together in the same village, the wives joining from elsewhere. Conversely, if goods passed down the female line, then it was often the husbands who joined their wives’ family villages. In essence, Murdock saw this as a way in which a community would maintain its accumulated stock. It was by no means a hard and fast rule. There were many exceptions and some completely different patterns of residence and inheritance. Nevertheless, there did appear to be a widespread trend, linking patrilineage and patrilocality on the one hand, and matrilineage and matrilocality on the other.

  Other anthropologists had seen the same trend as linked to different ways of life. Murdock quotes the words of an earlier anthropologist, R. C. Thurnwald:

  [S]ons inherit the trapping and hunting gear of their fathers; daughters the cooking utensils and fo
od gathering implements of their mothers. When the women have advanced from collecting to agriculture, their property is augmented, and matrilineal inheritance consequently becomes the more important.

  (Thurnwald, 1932:193-4, cited in Murdock, 1949:205)

  It was this kind of observation that led to the generalized idea of a shift from patrilineal inheritance and patrilocal residence among hunter-gatherers, to matrilineal, matrilocal ‘lower agricultural peoples’, changing back again in ‘advanced’ hierarchical societies. This generalized idea is heavily imbued with rather dated notions, such as a fixed notion of gender roles, and the unilinear evolution implicit in such words as ‘lower’, ‘advanced’ and so on. There was little archaeology could do with this idea beyond pasting kinship patterns observed today rather uncritically upon the mute remains from the past. There is still very little kinship evidence from ancient DNA to enable us to explore the issue, but enough to see how the work might unfold. Three of the most detailed studies to date are distributed rather arbitrarily around the world, but by chance they come from a hunter-gatherer community, an early farming society, and a relatively recent pre-Columbian community. Let us consider them in reverse order.

  Anne Stone and Mark Stoneking had gleaned several pieces of valuable genetic information from their pre-Columbian burial mound on the Illinois River at Norris Farm. Much of this information has contributed to the story of America’s first colonization, discussed in Chapter 7. This burial mound had been associated with a rather short-lived village of the Oneota culture, on a distant periphery of that culture’s central range. This is just the kind of context in which we might expect quite a small effective breeding population and a discernible pattern of kin relations. What the researchers actually encountered was a high degree of genetic diversity, and a marked absence of maternal or sibling patterns on the ground. This was, however, based on mitochondrial evidence alone.

  They amplified DNA sequence information from over 100 skeletons, and for more than fifty this included detailed information about the mitochondrial control region. For the control region section examined, three-quarters of the individuals carried rare or unique lineages, 50 percent more than the average rate in modern populations. No sign of a genetic narrowing there. Might men and women show different patterns? So far as Stone and Stoneking could see, they did not. The diversity was high in both sexes. What about the positions of the burials on the mound–might there be family clusters at different points? The major lineages displayed no such grouping. Each of the lineages was found across the mound. The cultural evidence may have been of a community that was remotely situated and perhaps isolated, but the mitochondrial picture was of considerable mixing, a true melting pot of maternal linages.

  This is the kind of pattern we might expect from patrilocality, or indeed one of the less common residence patterns, but certainly not from matrilocality. It is not a surprising result; many living Native American communities come from a patrilocal tradition. Where we might expect to encounter matrilocality is among the early agricultural communities of the Neolithic period. These have often been regarded as forming an egalitarian, matrilineal interlude between the early hunters and the later hierarchical warrior societies in which the male line came once more to the fore. The early farmers of Europe left many a collective burial across the landscape, often within a substantial monument of earth or stone, yielding ideal populations for kinship analysis. One such monument was excavated in Calvados in northern France, and scientists at the Institut Pasteur at Lille set about tackling the issue of the relationships between the buried individuals within.

  The burial monument constructed for the early prehistoric farmers at Conde-sur-Ifs was made up of six passage tombs, each with around a dozen bodies within, laid to rest 7,000 years ago. In one of these chambers, excavation revealed eleven skeletons laid out in such a way as to suggest two family groupings. Each grouping was made up of a cluster of bodies which included both adults and children. Thomas Delefosse and Catherine Hanni wondered whether they might be able to emulate the study of the Romanov remains, but on a much older set of bones. From the two putative family groupings they took samples from three adults and two children and set about amplifying and sequencing parts of the first and second hypervariable segments of the mitochondrial control region. The three potential outcomes they had in mind were: that both groups were closely related; that the groups corresponded to two unrelated families; or that the individuals in the chamber were not maternally related at all.

  What they found was that in the five individuals examined from the eleven bodies in the chamber, in relation to both hypervariable segments all their mitochondrial sequences were distinct. The individuals were clearly not maternally related. This is not to say they were not related at all. However, the mother-child relation that might have been inferred from the arrangement of burials was certainly an illusion. Moreover, they were not sibling groups. Despite the rather cosy grouping suggested by the six groups of around a dozen bodies each, themselves arranged in small clusters, the genetic diversity of those laid to rest within the tomb was clear. The authors admit that this is only the start of the work, but a start which already disposes of some assumptions. What is absent from the Conde-sur-lfs’ data is a reflection of matrilocality within the early farming community buried here. As with the much later Norris Farm community on the far side of the Atlantic, the mitochondrial diversity suggests something different. It will come as no surprise that the third assemblage, older still, also displayed significant mitochondrial diversity.

  Like the mound at Norris Farm, the ancient sinkhole at Windover in Florida has also figured several times in the short history of biomolecular archaeology. The bodies preserved within it, most of all their extraordinary pickled brains, are evidence of how abruptly decay can be arrested when certain conditions of water, air and acidity coincide. One such brain from the Windover pond yielded one of archaeology’s first findings of ancient human DNA. As the number of bodies studied increased, so the pond took on a new interest. It became the site where some of the rarer New World mitochondrial lineages, now almost confined to the heart of the Amazon, were widespread in prehistory. By the mid-1990s, over 170 individuals had been lifted by the archaeologists, and 50 percent retained soft tissue intact. Bill Hauswirth and his colleagues could now probe yet deeper into the communities that deposited their dead in the Windover pond 7,000-8,000 years ago. As with the other burial studies discussed above, the mitochondrial control region provided an obvious sequence for study, and Hauswirth and his colleagues amplified a 168-base-pair sequence within the control region. Among fourteen individuals sequenced, they found eleven distinct haplotypes, another very high proportion. The Windover population also yielded ancient DNA data so far unrecorded from the other two burial groups discussed above.

  The reader may by now be wondering why several of these kin studies have looked at the mitochondrial patterns alone. There would seem to be an obvious pairing between mitochondrial studies tracking female lines, and Y chromosome studies tracking male lines, and this will indeed form the basis of ancient kinship studies in the future. What we see in past work is the parallel unfolding of modern human genetics as a science. Information about the mitochondrial genome came on stream ahead of other parts of the genome, and our detailed knowledge of the Y chromosome has only recently expanded to its current state. Anne Stone is keen to get back to the Norris Farm individuals to examine Y sequences, but that is in the future and subject to the consent of the relevant living communities. The same is true of Windover Bog, but what that population has yielded is some autosomal information.

  In addition to examining mitochondrial sequences, Bill Hauswirth looked at other parts of the genome, less frequently targeted in ancient DNA analysis but dovetailing with a major strand of modern human genetic systems. One of the most valuable genetic markers among the blood proteins is a system of proteins on the surface of white blood cells, part of the body’s armoury against invading disease o
rganisms. These proteins, the human lymphocyte antigens, or HLA for short, are probably the most variable protein system within the body. Indeed, a variety of defence systems is central to ensuring that at least some people survive to reproduce however heavily disease hits. That great diversity has incidentally been of enormous value as a marker of human genetic diversity, and has been central to Cavalli-Sforza’s genetic synthesis. The DNA blueprint for HLA is to be found on an arm of the sixth chromosome in the nucleus as a string of genes, each one with ten to fifty possible alleles. Hauswirth decided to extend his study from the well-worn path of the mitochondrial control region to the nuclear genome, and to chromosome six. He picked a hypervariable, non-coding region within the HLA encoding arm of the chromosome, and charted its diversity within a sample of fourteen individuals. The diversity was high, another reflection of the mixed gene pool that first entered the New World, but one particular element caught his attention.

  At one specific locus, all except one of the individuals shared one allele in common. In many a less variable region this would not merit comment, but in this highly variable part of the genome, it did suggest a blood connection between the individuals. This possibility Hauswirth was able to explore by looking at some other highly variable regions. Like many interested in recovering ancient kin patterns, he turned to microsatellites. Hauswirth and his colleagues focused upon one particular gene on the first human chromosome. This gene contained one of those microsatellites that Hagelberg had exploited, made up of tandem repeats of the cytosine-adenine couplet. There may be fewer than ten, or more than twenty couplets in the microsatellite, according to genetic lineage. Among the Windover individuals, however, two particular microsatellite lengths dominate the population, one of nine repeats and the other of fifteen repeats. Alongside the HLA evidence this was further indication of close blood ties within a population placing its dead in the pond over several centuries.