When I was in the process of developing my course on human variation I discussed the course with my colleague and friend, Doug Crews, professor of anthropology at the Ohio State University.  He had been teaching a course in human variation and the two of us decided we would try to put together a reader that we could use for both our classes.  One of the articles I proposed was chapter VII of Darwin’s Descent of Man, the chapter entitled “On the Races of Man” (1871).  While there are many 19th century racial anachronisms in this chapter, in it Darwin makes a strong argument for the unity of the human species, all groups coming from a common ancestor, and for the differences being largely the result of adaptation to environmental circumstances.  For many years I required my students to read this chapter to contextualize the 19th century race concept and to see the argument Darwin made against polygenism or the separate origins of the races.  Graves, in The Emperor’s New Clothes which I have used in my race class, summarizes the critical paragraph from The Descent of Man in table form (2001:66).  This is the same paragraph to which many introductory anthropological texts call attention when discussing human variation.  In this selection, Darwin argues for the lack of clear boundaries between races and that therefore there is only a single human species with a single origin:

          ​I have used this statement in my presentations many times to help convince students of the social nature of the concept of race.  But something about it troubled me.  I knew that many anthropological works cite Buffon as contributing to the 18th century understanding of human variation.  His suggested means of adaptation leading to race formation within the human species includes ideas that are still being investigated today. As he notes:

          But I also had been teaching throughout my career that Buffon was the anti-Linnaeus, he opposed the idea of classification as a valid pursuit in natural history.  After using Jon Marks’ book Human Biodiversity (1995) as a required text in several courses, this idea was strongly reinforced by his extensive coverage of the contrasting approaches of Linnaeus versus Buffon in early anthropology.  This led me to question the place of Buffon in understanding the history of the concept of race in anthropology.  If Buffon was anti-classification, why would he divide the human species up into a small number of races like other 18th century scholars?  Why would he use the concept of race in this “modern” sense of a few large groups?

Uses of the term “race”
​
           The idea that Buffon brought the term race into scientific discourse goes well back in the anthropological literature.  Montagu (1942) blames Buffon for the term while arguing how dangerous it is.  Montagu understands, however, that Buffon’s use of the term “race” is nothing like the usage seen since the early 19th century:

          ​As a contrasting view, Nott and Gliddon (1857) assert that Buffon’s work was secondary and it was Cuvier who was truly responsible for the scientific classification of mankind:

​           Smedley (1996) cites Scheidt (1950) as her source for Buffon bringing race into the vocabulary of the natural sciences.  The Scheidt piece is actually a 1925 German language publication which was translated, edited, and reprinted by Count in his 1950 reader, This is Race.  I point this out because one of the problems with understanding Buffon’s influence on the concept of race has been the use of secondary sources, translations, and abridgements of his work that do not always clearly capture the intent of Buffon.  Scheidt adds in a footnote to his article, “the word ‘race’ to all appearance was introduced into the language of natural science by Buffon,” Scheidt (1950:360).  Smedley (1996) goes on to suggest that the term was used by other earlier workers but none as significant for subsequent natural historians as Buffon.
           While Buffon may have been responsible for influencing many other workers to use the term race for what had been called varieties of the human species, as Montagu clearly points out, he was not using it the way later workers used it.  Here are some examples of the use of race in Buffon’s 1749 of the Varieties of the Human Species:

           These examples, drawn from Barr’s translation of Buffon’s Histoire Naturelle illustrate even more clearly than Montagu’s that Buffon is using race in a 17th and early 18th century sense as synonymous with people or nation or society.  His usage is closer to what would be called an ethnic group today (Crews and Bindon, 1991).
           The last example also calls attention to another issue with Buffon’s discussion of the varieties of man and that is his uncritical acceptance of the existing literature for most of the groups that he discusses.  Gossett (1963) says that Buffon’s work was sometimes referred to by other authors as “unnatural history” making a play on the title of his huge volume of work to reflect the fact that Buffon’s credulity was greater than that of many of his successors and competitors.
           If we want to continue to credit Buffon with introducing the word race to the “scientific” literature, we need to do so the way that Montagu did, noting that it was a usage that would not be familiar to most readers today—or even through most of the 19th century.

How many races?

          Now for the idea that Buffon divided humanity up into six races; as noted by Montagu this is apocryphal, but it has had substantial staying power within the anthropological literature, from the 18th century up to the present. In a 2007 article in Annals of Human Biology, Biondi and Rickards repeat the six race orthodoxy, and the same year, Madrigal and Barbujani compile a table on the premise of Darwin’s paragraph and they also cite Buffon as having a six race system. The latter pair also attribute six races to Linnaeus (who never used the term “race”) by including his wild men and his monstrous men! In the 2006, 6th edition of Molnar’s Human Variation: Races, Types, and Ethnic Groups, he not only claims that Buffon had a six race system, he names the six races: Laplander, Tartar, South Asiatic, European, Ethiopian, American (Molnar, 2006:6). These six are the same ones listed by Madrigal and Barbujani and I suspect Molnar is the source for that part of their compilation.  Barbujani repeats this table in a 2010 publication with Colonna.  So this claim has continuing legs.
          In my presentations to various anthropology and biology classes over the years I have used variations on the table that Molnar presents the same way that I have used Darwin’s paragraph to illustrate the difference of opinion on how many races there are. This issue of Buffon’s six races was something that I just could not put out of my mind so I decided to try to find out how this got into the anthropological literature.
           I started with Darwin’s mention of Buffon’s six races in The Descent of Man.  After completing the list of numbers of races discussed by the various workers, Darwin has a footnote that reads as follows: “See a good discussion on this subject in Waitz, Introduction to Anthropology, Eng. translat., 1863, pp. 198-208, 227. I have taken some of the above statements from H. Tuttle’s Origin and Antiquity of Physical Man, Boston, 1866, p. 35.”  Here is the statement of Tuttle’s to which Darwin refers:

​           As we can see, Darwin took quite a lot of Tuttle’s treatment for his own use.  The preceding paragraph in the two pieces is remarkably similar.  It is interesting to note that the list of races that Tuttle lists for Buffon does not include any race from Africa.  Since other writers had taken issue with Buffon’s disdain for Africans before Tuttle, it is surprising that the absence of Africans did not raise an alarm.  The other source that Darwin lists, Waitz, reviews Buffon’s work in great detail and relies on many of his biological and ethnological statements, but never suggests that Buffon classified humans into a discrete number of races.  Once I had looked up Tuttle’s confusing statement and Waitz’s (1863) fairly clear treatment of Buffon, I was at a dead end because Tuttle cites no source for the six races of Buffon and Waitz does not mention them, so I began looking for other authors who could lead me to a source.  I quickly found Brewer’s statement about Buffon:

​           Already we can see that the races are very different in the version given by Tuttle and the one from Brewer and neither one exactly corresponds to the list given by Molnar (derived from Hrdlička, but Molnar cites Slotkin, 1965 and Montagu 1960).  Like Tuttle, Brewer names no source for his races of Buffon, although it is “now very generally accepted” (Brewer, 1890:117).
           The one Buffon aficionado who actually espouses a six race system for man is Goldsmith ([1774] 1854).  Much of Goldsmith’s 1774 natural history is based on the work of Buffon, but when he comes to humans he takes a very interesting tack giving his own list of races:

​           We see here that the explanation for the origin of these races is essentially the same as that espoused by Buffon, a single human origin with adaptation to different environments accounting for the races.  The six races enumerated by Goldsmith are: Laplanders, Tartars, Southern Asiatics, Africans, Americans, Europeans, essentially the same as Molnar’s list, although Goldsmith doesn’t actually label the races the way Molnar suggests Buffon did.  Goldsmith offers the following note after saying that humanity is divided into six varieties: “I have taken four of these varieties from Linnaeus, those of the Laplanders and Tartars from Mr. Buffon,” (Goldsmith [1774] 1854:209).  Slotkin (1965) cites this note, attributing the six race system to Goldsmith, but it appears that few have bothered to take notice of it in the continuing attribution of this system to Buffon.
           In my search for the six races of Buffon, two authors cited sources for the number and name of the races: Molnar (2006) and Hrdlička (1941).  Molnar lists Buffon’s six races as part of a table, comparing the four races of Linnaeus, six of Buffon, five of Blumenbach, and three of Cuvier.  He notes at the bottom of the table:

​About this table of races, Linnaeus, Buffon, and Cuvier emphasized behavioral and cultural characteristics in their discussion of the varieties of man, Blumenbach less so. Following up on Molnar’s sources, I checked Slotkin (1965) and found a very clear statement that Buffon did not divide humanity up into races, and Slotkin took notice of the note from Goldsmith about the six race system. Rather than indicating that Buffon classified races, Slotkin provides lengthy quotations to illustrate the non-modern use of the word race by Buffon. Montagu, on the basis of his 1942 statement, was out of the question as a source for Buffon’s six races, but I checked the 1960 text anyway and found no statement at all about Buffon and race. Buffon was discussed exclusively in the context of the history of the concept of evolution.
           When I read what Montagu had to say about Buffon in his 1942 book, I followed up on a footnote about the six race classification and was led to Hrdlička’s 1941 piece on human races.  Hrdlička not only lists the races, he presents them in a geometrically arrayed fashion as so (1941:174):

​While Buffon offered very unflattering descriptions of almost everyone but the French, he never came close to enumerating these six races and he never attempted to arrange them in some geometric or geographic system as Hrdlička implies.  Furthermore, the choice of the term Ethiopic is especially enlightening about the care that Hrdlička took in his scholarship of Buffon since Buffon himself offers this statement, “By confounding the Ethiopians with their neighbours [sic] the Nubians, who are nevertheless of a different race, we have been long in an error with respect to their colour [sic] and features,” (Buffon, [1749-1788] 1792:272).  He also says, “These contrarieties are more than sufficient to confirm us in the opinion that the Hottentots are of a race distinct from that of the Negroes,” (Buffon, [1749-1788] 1792:294). He clearly had no intention of classifying all African or sub-Saharan African peoples in a single race.  Hrdlička cites Buffon’s Methode en histoire naturelle; la theorie de la terre et de l’homme, 1749-1789, as his source for the six race system and list.  A search of the Histoire Naturelle web site (http://www.buffon.cnrs.fr/) provides no such title and a comprehensive Buffon bibliography similarly has no such title by Buffon (Genet-Varcin and Jacques Roger); it appears to be a scrambled version of titles in the series of Histoire Naturelle, but it is clear from what Hrdlička presents that he has not consulted Buffon.  Perhaps he mistook Goldsmith’s list of races for Buffon’s he might have consulted a translation by John Wright (1831) that includes copious notes inserted by the translator including the ideas of other authors such as Linnaeus.  Wright goes on at some length about the classification of Bory de Saint-Vincent and others in the middle of the translation in a way that might mislead the incautious reader.  In any event, if Hrdlička did in fact research Buffon, he must have used some 19th century source that scrambled Buffon’s intent, but no translation nor original of Buffon’s work bears the title that Hrdlička cites or presents the six race system that he indicates and the geometric arrangement is pure fancy.
           Buffon’s work was very influential and it inspired numerous translations into many European languages.  Not all of the translations sought to directly recreate Buffon’s thinking intact.  An 1800 translation published in Edinburgh abridges the 36 volumes of Buffon into two volumes.  On the title page the abridgers note that the work has been compiled chiefly from Swammerdam, Brookes, Goldsmith, etc.  This volume offers the following leader to Chapter VI: “of the apparent varieties in the human species—Laplanders—Tartars—Chinese—Japanese—Formosans—Moguls—Persians—Arabians—Circassians—Turks—Russians—Negroes—Hottentots—Americans—causes of this variety,” (Buffon, 1800:48).  Here the abridgers list 14 races, not including any mainstream Europeans.  This is more indicative of the way that Buffon uses the term race, but even this list is clearly not an exhaustive catalog of Buffon’s races. This chapter leading statement is not a translation from Buffon, as there is no such header to his Volume IV, Chapter IX, “of the Varieties in the Human Species.” Even these abridgers, drawing from Goldsmith as they are, do not suggest a six race system of classification for Buffon.
           I recently found a reference to Buffon’s races that took me into the legal literature.  Reading Dorothy Roberts’ 2011 book on race, I came across a mention of Buffon in reference to the decision against the naturalization of Chinese in the 1878 case In re Ah Yup.  Roberts says the California state jurist, Judge L.S.B. Sawyer, after reciting precedent saying that “white person” referred to someone of the Caucasian race, “then turned for guidance to the racial typologies developed by European naturalists Johann Friedrich Blumenbach, George Louis de Buffon, Carl Linnaeus, and Georges Cuvier, observing that all grouped Caucasians separately from Mongolians” (2011:15).  This led me to the text of Sawyer’s decision where he stated that “[Blumenbach’s five race system] was adopted from Buffon, with some changes in names, and is founded on the combined characteristics of complexion, hair and skull.”  This revelation of a five race system for Buffon was the first time I came across that configuration and the idea that Blumenbach borrowed his system of racial classification from Buffon was completely surprising and baffling.  Sawyer cited the “Ethnology” entry of the “New American Cyclopædia” (Ripley and Dana, 1859).  This source has several pages discussing different racial classification schemes and has this to say about Buffon and Blumenbach:

Needless to say, there is no more validity in attributing this racial classification to Buffon then there is for any of the other works. Ripley and Dana refer readers to the notorious Types of Mankind (Nott and Glidden, 1854) as their source.
           Josiah Nott is responsible for much of the text of Part I in Types of Mankind, including the section where the history of the classification of human races is discussed (Nott and Glidden, 1854:xiii-xiv). Nott says on page 82, “Buffon divides the human race into six varieties—viz., Polar, Tartar, Austral-Asiatic, European, Negro, and American.” There are two noticeable errors in this statement. First, Buffon was actually quite careful with his definition of species and used the word race—not variety—deliberately to combat Linnaeus’s classification system (Hudson, 1996:253). Second, the six races are not consistent with Buffon’s writings, and I can’t figure out where Nott took this interpretation from. Apparently he was tutored by Louis Aggasiz in the science that was to go into Part I, so he may have gotten this list from him (Horsman, 1987:177-178). The execrable bias and scrambled scholarship of Types of Mankind notwithstanding, this book was the 19th century bible on race when it came out. The fact that evolutionary theory theoretically threw out the notion of polygenesis only five years later had very little effect on the impact of the book. The first edition was sold out before it was printed and the subscribers included Secretary of State Edward Everett, who also purchased a copy for the State Department; the Treasury Department; and Secretary of the Navy John Kennedy purchased a personal copy and one for the Navy Library. By 1871 the book had gone through 10 printings with tens of thousands of copies distributed. It is clear that the book was still exerting influence in Judge Sawyer’s 1878 decision, and it is likely that the idea of Buffon’s six races (ignoring the error in the 1859 Cyclopedia) became so widespread in the scientific literature because of its citation in Types of Mankind. Of course, later in the century people would not want to refer back to it, so when Tuttle, writing 12 years later—post Origin of Species and post-Civil War—in Boston, scrambled the races but kept the number suggested by Nott he also borrowed several of the other classifications that Nott included in that section, without attribution.
           As Montagu pointed out in 1942, Buffon never suggested a six race classification of humans and those who suggest he did cannot agree on what the races are nor can they point to a source in Buffon’s work that would validate this claim.  The fact that Darwin, Hrdlička, and more recent biological anthropologists make this mistake emphasizes the importance of finally getting it right.

Conclusion

           So what are the lessons to take away from this exercise? First, attributing the use of the word “race” to Buffon is partially valid, but if you want to make this claim it needs to be qualified as Montagu (1942; 1996) did or contrasted with the Linnaean perspective as Marks (1995) did.  Buffon may be responsible for introducing race into the scientific literature on man, but it was a race concept that has very little to do with the modern concept.  Second, Buffon never classified humans into six races.  The fact that so many formidable anthropologists maintained this apocrypha which is probably derived from the reviled Types of Mankind is testimony to the staying power of academic falsehoods.  However this myth originally crept into the anthropological literature, and my money is on misreading Goldsmith and attributing the system to Buffon, it has been almost impossible to eradicate right into the 21st century.
​           The teaching moment from this project is to not believe what anyone says about anyone else’s work.  In teaching research writing to students for many years, I emphasized the dangers of secondary sources and instruct students to go to the original sources (Bindon, n.d.).  I discovered how easy it is for apocrypha to creep into a reference list when I cited one of my very early papers in another one I was writing, only to find out between submission and revision that I did not actually make the case I was citing the paper for in that particular publication.  If I could mistakenly cite my own work, think what I could do to the work of others!  The case of Darwin’s allegory and Buffon’s apocryphal races provides a good example of the dangers of relying on the scholarship of others.