Chadic languages and Y haplogroups.

The January 2009 publication in this journal of an article entitled ‘Human Y chromosome haplogroup R-V88: a paternal genetic record of early mid Holocene trans-Saharan connections and the spread of Chadic languages', by Cruciani et al,1 represents a major step forward in our understanding of the African Y haplogroup diversity and pre-history. Approximately one year ago I published a detailed review of the earlier studies in this field.2 Similar to Cruciani et al, a major theme of my consideration was the probable link between the dispersal patterns of Y haplogroups and Afro-Asiatic languages. In particular, although the major focus of that review was upon Y haplogroup E1b1b1 (E-M35), similar to Cruciani et al, I suggested that in the specific case of speakers of the Chadic family of languages, the high frequency of R haplotypes, which are otherwise uncommon in most of Africa, seems to be relevant for judging the strength of the competing theories regarding the origins of this language group within Afroasiatic. A comparison with this new publication can help put its findings into useful perspective. In my review I proposed that the Y chromosomal evidence seemed to be most compatible with Blench's3 theory that Chadic is ancestrally most closely related to Beja and Cushitic, and arrived in its current position from the east and the direction of Sudan and the Red Sea. At that time, the data seemed to be much less compatible with Ehret's more novel position that it arrived from the north and is most closely related to Berber.4 A key consideration is that R haplotypes in various forms are generally Eurasian, being found as far away as Siberia, and their concentrated presence this deep in Africa has long been considered as evidence of movement of people from the direction of Levant, although the timing and route have been difficult to determine.5, 6 Blench's proposal that Chadic arrived with pastoralism involving goats, a way of life that spread from the Middle East, probably initially along the eastern side of the Nile, therefore matches the DNA evidence very closely. Furthermore, I observed, as do Cruciani et al, that the discovery by Cerny et al of a mitochondrial link between Chadic speakers and the Horn of Africa (Mitochondrial haplogroup L3f) also supports Blench's position.7 In summary, I proposed that, in this case, a Y haplogroup had possibly been carried by a wave of innovation in life style (pastoralism), independently of language or overall genotype. As I remarked: The recent autosomal DNA study of Tishkoff et al. (2009)8 confirms that on the whole, Chadic speakers are more closely related to their Nilo-Saharan neighbours than to any other Afroasiatic group. Looking at the genome beyond Y-DNA these peoples show far less Eurasian ancestry that the Beja, for example. Nevertheless, Cruciani et al argue that the discovery of V88 now demonstrates a strong Y chromosomal link between Chadic speakers and speakers of other Afro-Asiatic languages to the north of the Sahara. Their key evidence for this is their new data for Siwa in Western Egypt, a Berber-speaking area having approximately 27% R-V88 (93 people tested). These data are surprising and should be seen as one of the major discoveries of the article, and requiring an explanation such as the one given by the authors. The article contains a striking contour map, which shows a coloured band of R-V88 passing from Siwa down to the Chadic speaking area, which would match Ehret's proposed migration route. However, analysis of the article shows that this coloured band simply joins two areas with high frequency, Siwa and the Chadic area. There are of course no data for populations along this band, which runs through the Sahara. One single population is therefore very important in their account. The potential importance of gaps in the data should be considered in several other directions also. The study also contains no data for Sudan, which lies between the Chadic speakers in the west, and the Afro-Asiatic speakers in the horn of Africa and near the Red Sea. There are also no data for eastern Egypt. These areas are critical in determining whether Blench is likely to be correct, because they represent the path along which he believes pastoralism spread. Data for Sudan are limited but, as the authors note, the 2008 paper of Hassan et al does seem to indicate a potential presence of R-V88 there.9 We know from this and other studies that E1b1b1 dominates the Berber speaking area, leaving the Siwa result as quite surprising and unique. This in turn raises the question of what other surprising and unique populations may exist in other oases and unsampled regions. With the limitations of the data in mind, it could perhaps be argued that Cruciani et al's hypothesis is strengthened by the similarity between their age estimate for the R-V88 clade and the time depth at which we know the Sahara was green, making movement in the area between Siwa and Lake Chad more practical. However, unfortunately, such age estimates are one of the most controversial parts of any study of this kind. Different estimation techniques are likely to give a wide range of possible answers.