When an anthropologist or geneticist wants to look at historical populations, surnames provide a stand-in for genes - in our society, names are inherited patrilineally, that is, from father to son, like a sex-linked gene. It is difficult to collect DNA samples from people who died two hundred years ago; it is very easy to collect the surnames of an entire population of that age. So isonymy, the study of surname patterning, can provide a lot of information about the population biology of historic human groups.
This coming week I am presenting a poster in San Antonio at the annual meeting of the American Association of Physical Anthropologists on the patterning of surnames in 18th century Brooklyn, New York. Why surnames ? Because when an anthropologist or geneticist wants to look at historical populations, surnames provide a stand-in for genes - in our society, names are inherited patrilineally, that is, from father to son, like a sex-linked gene. It is difficult to collect DNA samples from people who died two hundred years ago; it is very easy to collect the surnames of an entire population of that age. So isonymy, the study of surname patterning, can provide a lot of information about the population biology of historic human groups.
In such studies, one calculates various genetic statistics, including the degree of similarity between different human groups, or populations, and the degree to which each such population is a homogeneous whole in and of itself. So in the study that I am about to present, I found that the six Dutch towns in 18th century Kings County remained very isolated from the outside world despite the fact that they were immediately across the river from the bustling metropolis of New York City. These statistics work just fine, if we assume that all surnames truly are inherited father-son. But are they? There are two problems with this assumption. One is pholyphyly, the fact that some, if not many, names can be traced back to multiple origins. The second is what geneticists politely call non-paternity: The fact that you bear your mother's husband's name does not ensure that he is in fact your father.
Surnames are not the only things inherited patrilineally. So are our y chromosomes: Since females have two x chromosomes, while males have one x and one y, your y chromosome (or your father's, if you don't have one yourself) can be traced back through a never-ending chain of fathers. In this respect, studies of y chromosomes complement those of mtDNA which is inherited matrilineally. Y chromosome analyses have previously proven useful in genealogical research when paternity was in question, as in the case of the African-American descendants of Thomas Jefferson. They have also painted a slightly different picture of relationships between populations on a global scale than mtDNA has.
In the April 2000 issue of the American Journal of Human Genetics, Brian Sykes and Catherine Irven of the Institute for Molecular Medicine at Oxford University report the results of a study they conducted of the y chromosomes carried by British men named Sykes. Like most British surnames, Sykes can be traced back to one region of the country, but it has usually been assumed that such names had originally been adopted by several men in that area around the same time, in the 13th and 14th centuries. When they examined y chromosomes from 48 Sykes men, they found that 43.8% carried chromosomes that were either identical or very closely related, and were thus descended from one man, presumably the first to bear the surname. The others bore a completely random assortment of chromosomes. This indicates that there was probably no other original Sykes, at least not one who left many descendants. Instead, the other men have at least one incident of non-paternity somewhere in their genealogy. This does not necessarily mean illegitimacy or adultery; rather, it could be the result of adoption. Without detailed genealogical work, we cannot differentiate between these causes.
This percentage of non-paternity over the seven centuries since the formation of the surname only requires 1.3% of the children in each generation to be the children of different fathers. This rate is actually quite low from a comparative perspective: other studies of European and American populations have estimated rates of over 10%.
Like many other British surnames, Sykes was carried over the Atlantic to North America by at least one immigrant - although we do not know if he bore the same y chromosome as the founder of the family. Its spread over the next few centuries illustrates how individual families have spread and moved across the continent. It is quite possible, although untested, that none of the American Sykes bear the same y chromosome as the British ones. This is due to what geneticists call "founder effect." Most surnames were carried over the Atlantic by a handful of people, sometimes only one. These migrants were a very small sample of the British population. If only 60% of the British Sykes bore this particular chromosome in the 17th century, and only one migrated, there is a 40% chance that he had a different chromosome. It will be interesting to see what future studies of this and other surnames in both Old and New Worlds find.
From my personal perspective, this study is particularly important because it provides support for the underlying assumptions of my surname work: Many surnames are monophyletic, that is, descended from a single individual, and non-paternity rates are not extremely high.
So names can tell us a lot about historical population genetics. But what about their social significance ? In a future feature I shall examine this question: Why do different cultures choose different ways of naming their members, and what can such patterns tell us about a people ?