Dr. Eugene Foster
Former Professor of Pathology,
University of Virginia
"The Intersection of Science and History: DNA and the
Thomas Jefferson and Sally Hemings Story"
September 27, 2000 

Dr. Foster: It’s been almost three years since my collaborators and I published the results of our DNA study of the paternity of Sally Hemings’ children. So of you may remember the voluminous and extraordinary reactions to the few bits of new information that we provided in our 729-word article. As you know there is still a lot of controversy about the significance of the DNA findings. After many months of considering the historical evidence seen in the light of the DNA findings, a special committee at Monticello, at the Thomas Jefferson Memorial Foundation, composed mostly of historians, concluded that it was most likely that Jefferson was the father of at least one of Sally Hemings’ children, Eston Hemings , and may have fathered all of them. One member of the group, however, dissented and another committee, not affiliated with Monticello, is reexamining the whole subject and expects to issue a report in January. So it’s clear that the DNA findings have not resolved all aspects of the controversy. I hope to explain why this is the case.

Let me start by telling you how I happened to do the study, and "happened" is the right word. I’m not very qualified either by education or experience to have done anything like this. I’m not a molecular geneticist; I’m not a historian; and I’m not a genealogist. During most of my career as a professor of pathology, I taught medical students about the causes, processes, and manifestations of disease. I did autopsies; I made diagnoses on biopsies; I did research on various types of cancer and bacterial infections. None of these activities involve much knowledge of DNA. And although I was aware of the allegation that Thomas Jefferson might have fathered Sally Hemings’ children, I had no great interest in the subject. But in the spring of 1996, Winnifred Bennet, a friend whose curiosity had been peaked by the use of DNA in the Anastasia affair, that some of you may know about, asked me whether I thought DNA could be used to resolve the Hemings-Jefferson controversy. I told her I really didn’t know but I’d try to find out.

On and off during the next year I learned how DNA was being used to identify individuals and to determine paternity and I also read something about the controversy and about the genealogy of the families involved--the Jeffersons, the Hemings, the Carrs, and the Woodsons. By the spring of 1997, I had concluded that although it was theoretically possible for DNA studies to provide some answers, the probability of success was so slight and the effort and expense would have been so great, that it wouldn’t be worth trying. And that shouldn’t have been surprising because the idea had been around for a long time but nobody had done it.

Well, how did I reach this conclusion? Now I have to get just a little bit technical, a little bit of genetics and DNA. This is what a set of chromosomes from a man looks like and most of our DNA is contained in these 23 pairs of chromosomes. You’ll notice that they’re arranged in pairs. One of each pair comes from our mother and one from our father. So half of our DNA comes from our father and half comes from our mother. Before we leave this light, let me also point out the last pair of chromosomes way down at the bottom, the X and Y. And I’m sure as you know that the big one is called X, the small one is called Y. And this set of chromosomes is from a man because it has one X and one Y. A woman has two X chromosomes. So you get your X chromosome either from your mother or your father, but a man gets his Y and the DNA in it only from his father.

Now how is DNA used for paternity testing? Using one or more complicated tests, it’s possible to show that there are over 200 locations in the DNA, which vary a great deal from person to person. There is so much variability that the chances of one person having the same combination of markers in these areas is less than 1 in 30 million. So it’s possible to develop a so-called DNA fingerprint for a person, which can be used to identify him or her.

Now what about paternity testing? To determine whether a man is a child’s father, one does a DNA fingerprint of the man, of the child, and usually on the mother. Because the child has gotten about half of its DNA from each parent, its DNA should have about half the father’s markers and half the mother’s. So it’s pretty simple. So if this technology had been available when Thomas Jefferson and Sally Hemings’ children had been alive, and Thomas Jefferson and Sally Hemings were alive, it would have been easy to settle the question. But of course they’re not around now, so what about getting DNA from their bodies? All we’d have to do is exhume Jefferson’s body and those of Sally Hemings and her children. Unfortunately, except for Jefferson, we don’t know where the graves are, and we didn’t think we would have much chance of getting permission to exhume Jefferson’s body. And there are other questions also, like ‘did the DNA survive?’ and, ‘in fact, are the remains in Jefferson’s grave really his?’ You never know. Funny things have happened in the course of history. So we were left with the possibility then of trying to find living Jefferson descendants and living descendants of the other people involved in the controversy, and comparing their DNA fingerprints to see if there were enough similarities to make some statement about the paternity. Well that also turns out to be a real problem. Now this slide illustrates that. If we start out, let’s say that "all" represents Jefferson’s DNA. Now his children have half, actually less than half. The next generation a quarter, down the line. By the time you get to the fifth, sixth, and seventh generation there just isn’t much left. So that’s the main reason why we thought that DNA couldn’t really be used for this. In addition to that, some of the other people involved were likely to have had some of the same DNA as Jefferson--his nephew certainly. And so the problem just got immensely complicated and it would have taken literally thousands of DNA tests and very complicated statistical analyses and probably would not have really led any place at all.

But just when it looked as if there was no feasible way to answer any of the questions with DNA fingerprinting, Professor Rolph Bessinger, a member of our own biology department, accidentally learned of my interest in the problem and told me of some new knowledge about the Y chromosome that would make it useful for tracing paternity over many generations. He told me that Y chromosomal DNA had been found to have a moderate amount of variability. Previously it had been thought to be almost the same from one man to another…very, very few and small variations. But there were some new knowledge. What makes the Y chromosome good for this kind of study? It’s present only in man, as we have said. It passes unchanged from father to son, generation to generation. Father, son, son, down the line. And it does not characterize individuals but it does characterize families. Now ordinarily we would say this is a weakness, so if you want to do paternity testing, you don’t do it with a Y chromosome because a man has the same Y chromosome as his brother so how do we know who it is? Often the discussion is that way. So we are not talking about using it in the first generation. We are talking about using it in remote generations, all the way down the line and Y chromosomes can be traced in population studies, for instance, all the way down unchanged for dozens of generations.

Well if one could characterize Jefferson’s chromosomal DNA, it could be compared in the DNA of a man who claimed to be his descendant. NO because Jefferson had no legitimate line of descent, his only son had died in infancy, I had to find living men who were descended from one or more of Jefferson’s paternal relatives. A brother or a paternal uncle. And these had to be people who were descended in an unbroken male line, all the way down the line. So basically they had to be named Jefferson, unless they changed their names. So we were looking and it turned out that there were a number of people living not too far from here who were descendants of Field Jefferson, who was Jefferson’s father’s brother, Jefferson’s paternal uncle. And according to everything we know about the Y chromosome, he should have had the same Y chromosome as Thomas Jefferson. And so then if we could find similar male line descendants of the Carr brothers, or relatives of the Carr brothers (male relatives), we could do the same thing for the Woodsons and for the descendants of Sally Hemings’ later children, we might get some useful information. It’s not easy to find male line descendants in unbroken lines so many generations later. As we all know from our own families, male lines die out all the time. A man marries and has two daughters, or he has no children, or he doesn’t get married and that’s the end of his line. So we were lucky that with the voluntary help of a number of amateur genealogists and family members—Herbert Barger, Karen Carr Neddleton (he lives here in Charlottesville), Byron Woodson,-- the appropriate people were identified. So then I had to contact them to see if they would agree to participate in the study by giving me a little blood. This was a very discouraging process at first. A month after I wrote these various men explaining the project and asking for their participation, I had gotten only one reply. That man said he’d be honored to participate. But when I phoned the people who hadn’t responded to my letter, I was pleased and amazed that almost all of them said that they’d be happy to participate. And later when my wife and I went to their homes, from Philadelphia to Hilton Head, from Richmond to Columbus, Ohio, they welcomed us warmly, sometimes with brownies or cookies. After some conversation they stuck out their arms and let me dray their blood. Not one of them asked me for any sort of credential, not even a driver’s license. I came prepared with diplomas and documents of all kinds and nobody wanted to look at them, so I didn’t offer them.

Well as I said earlier, I’m not a molecular geneticist. How did I get the DNA analyses done? Soon after Professor Benzinger told me about the Y chromosome, I read as much as I could about it and became acquainted with the work of Dr. Chris Tyler Smith at the University of Oxford. It was plain that he was one of the leaders in the field, so I emailed him to ask whether he thought the project was feasible and also to recommend someone who might be willing to do the analyses. He emailed back that he would be happy to do them and that he would also try to enlist some other investigators working in the field. Eventually the analyses were done using three different methods, and this is of scientific interest, at laboratories of the universities of Oxfod, Lester, and Liden in the Netherlands.

Now let’s get more specific about the historical controversy. Here are the people. So we’ve talked about all of them except Martha Jefferson, and it’s not important to say very much except that it was widely believed and generally accepted that Sally Hemings had the same father as Martha Wayles, but not the same mother. Thomas Woodson is the man who is believed by the Woodson family and by some historians to be the first son of Sally Hemings, conceived in France and born at Monticello, and then later came under the care of the Woodson family and later gave rise to the whole big African American family. Madison Hemings was the next to the last son of Sally Hemings and Eston Hemings. Now here are the specific questions that we were trying to get some information about:

-Was Thomas Hefferson the father of Thomas Woodson?

-Was he the father of a child Sally Hemings conceived in France? (That’s part of the same question, but those are, in a way, two different questions.)

-Was Thomas Woodson the child Sally Hemings conceived in France? (All of these three things are questions.)

-Was Thomas Jefferson the father of any of Sally Hemings’ later children? (The later children were Beverly, a man who was allowed to run away from Monticello and disappeared into White society; Harriet, another child who was helped to run away from Monticello and disappeared into White society; and then Madison and Eston. Madison and Eston were freed in Thomas Jefferson’s will, lived with Sally Hemings for another six or seven of years in Charlottesville, and after she died moved to the region of Chilecothe, Ohio.)

Well, because we couldn’t approach these questions directly, we had to try to get at them indirectly through examination of living male line relative of these historic characters. So the questions that very specifically…the contemporary questions that we had to answer were these:

-Was Thomas Jefferson the ancestor of Thomas Woodson’s living descendants? ( If he was, then it was very likely that Thomas Woodson was Thomas Jefferson’s son.)

-Was Thomas Jefferson the ancestor of Eston Hemings’ living descendants? (If he was, then it’s likely, almost certain, that Eston Hemings was the son of Thomas Jefferson. Or, as the althernative, both of those were probably not going to be true although they could have been. Jefferson could have been the father of one or more of the children or one or the other of the Carr brothers could have been fathers of the other children)

[-Were Samuel or Peter Carr the ancestor of Eston Hemings’ living descendants?]

So at any rate, these are the questions that we could approach. I want to point out that the primary purpose of the study was not to get information that helped to identify living descendants of Thomas Jefferson. That’s become a big thing since the study, but that isn’t what we were interested in. We were interested in the historical questions and this was an indirect way of getting at it.

Now when we began the study, we knew that no matter what the results were, they could not provide definitive answers for any of the historical questions. We could have gotten results that didn’t say very much at all. Nothing might have matched. Now this somewhat complicated slide summarizes the important findings. Let me just point out a couple of things. Here, this Thomas is not President Thomas Jefferson, this is his grandfather. And here is President Thomas Jefferson down here. His father, Peter, and his uncle, Field. And Sally Hemings here. I did not make this chart. This comes from somebody else’s paper. I would not have put this down as a certainty, but at any rate, here is Thomas Jefferson and Sally Hemings. And the people we studied were all of these people at the bottom with various colored blocks. We know who the intervening people were. And the important thing is that here we have five descendants of Thomas Jefferson’s paternal uncle, Field Jefferson, all of whom have exactly the same Y chromosomal DNA type, with one exception—this man had one minor mutation, something that we would expect after this many generations. Not surprising. We had one, and one only, descendant of Eston Hemings, down here. There are two other living descendants of Eston Hemings who are incommunicado. There family and nobody else can be in touch with them. And we would have been delighted if we could have gotten blood from them, but we couldn’t. So we had to stick with this. These are the descendants of Thomas Woodson. As you can see, they are descended from two of his sons. Since that time we’ve had another descendant of a third one of his sons who has the same Y chromosomal type. This one you see differs from the others. What we interpret that as meaning is that either there was illegitimacy here or unreported adoption. It is a totally different type that has nothing whatsoever to do with the others. But he fact that in this case we have identified that all of these people have the same type is very, very strong evide nce that in fact this came from one person. And the same thing down here. Four, five out of…counting the other one…five out of six being the same is very strong evidence. And then here is the stuff for the Carrs. This is a descendant of Dabney Carr, who was the father of Peter and Samuel Carr and then these were some paternal uncles. So the point to be seen from this slide is simply that the only match with the Jefferson family Y chromosome is that of the descendant of Eston Hemings, a man named John Weeks Jefferson.

So what are the strictly scientific conclusions that we can draw from these data. Strictly scientific. I’m not talking about history or anything. Just what can we say from the science alone? Well we can say that it is very likely that some Jefferson, including Thomas Jefferson himself, was Eston Hemings’ father. Second, we can say that it is very unlikely that Thomas Jefferson or any other Jefferson was Thomas Woodson’s father. Third, it is very unlikely that a Carr was Eston Hemings’ or Thomas Woodson’s father. Are other strictly scientific interpretations possible? Yes, there are many other possibilities, but we believe they are much, much less probable.

Now the scientific findings, taken together with the historical evidence, led us to the following comment at the end of our article. "The simplest and most probable explanation for our findings are that Thomas Jefferson, rather than one of the Carr brothers, was the father of Eston Hemings Jefferson (which is the name that he took) and that Thomas Woodson was not Thomas Jefferson’s son." But these are not proved. For instance, just to give you an example of why we can’t even say what the probabilities that this is right are. The whole study, the entire study, is based on an assumption. The assumption is that Thomas Jefferson and Field Jefferson had the same Y chromosome. Now did they have to have the same Y chromosome? We believe that they did. But if Thomas Jefferson in fact was not legitimate, if his father was not Peter Jefferson…we have no way of knowing that. No one has ever said anything about it, but we don’t know that…then this whole study is down the drain. Or if Thomas Jefferson was the legitimate son of Peter Jefferson, so we know that Peter Jefferson and Field Jefferson had the same father? We don’t know that either. We can’t know it. And we can’t assign any number to those possibilities. So that’s why, in our statements, we use probabilistic terms, because we’re not sure. We think that the possibility that Thomas Jefferson was illegitimate is very small. We think that the probability that Peter Jefferson and Field Jefferson did not have the same biological father is very small, but it can’t be completely excluded. Two years later, having heard many opinions that went too far or not far enough, we still think our conclusion is valid. Are other interpretations of the data, both the historical and the DNA data taken together, possible? Of course. If objective reproducible scientific findings can be explained in a variety of ways, it’s plain that historical data, which is much less objective, is even more open to different interpretations.

So the controversy goes on and probably will do so for the foreseeable future. My colleagues and I believe that with the presently available information, the conclusion that we drew is entirely valid.