Update to Chikhi et al.'s "Clinal Variation in the Nuclear DNA of Europeans" (1998):Genetic Data and Storytelling—From Archaeogenetics to Astrologenetics? Lounès Chikhi Keywords Clinal Patterns, Spatial Autocorrelation Analysis, Neolithic Transition, Phylogeography When "Clinal Variation in the Nuclear DNA of Europeans" was published in 1998, its message was relatively simple and pretty clear from its title. First, we were using genetic data from present-day European populations to describe how this diversity was distributed, by way of spatial autocorrelation analysis. Second, we observed clear clinal patterns in most genetic markers; that is, we found that autocorrelation was positive and significant at short distances and negative and significant at larger distances. Third, the difference from previous studies was that the markers we used were directly based on DNA information. Sure enough there had been studies describing the spatial patterns of genetic diversity in Europe, but they were based on so-called classical markers (allozymes and blood groups). At the time the only studies using "proper DNA" were those based on mitochondrial DNA. Looking back, our paper came, to a large extent by chance, at an interesting moment. Indeed, for the last two decades studies by Cavalli-Sforza, Sokal, Barbujani, Oden, and others had shown that many "classical markers" were exhibiting Europe-wide clines (Cavalli-Sforza et al. 1993; Menozzi et al. 1978; Sokal et al. 1991). Although such clines could have been produced by selection, as a result of climatic gradients, for instance, it seemed unlikely to most population geneticists that selection could have acted in the same way on so many independent loci and over such distances. Simulation studies performed between the 1970s and the 1990s suggested that a major continent-scale demographic event had to be responsible for these clines (e.g., Barbujani et al. 1995; Sokal et al. 1991, 1996). Demographic events important enough to explain Europe-wide patterns were not many. Based on archaeological data, only two candidates seemed to fit. The first was the colonization of Europe by modern humans during the Paleolithic. The second was one of the most important cultural and demographic events in human prehistory (e.g., Mithen 2007), the Neolithic transition, during which human societies moved from hunting and gathering to farming and domestication. [End Page 639] In the early and mid-1990s, most people were probably accepting Ammerman and Cavalli-Sforza's (1973) model of demic diffusion against a model of cultural diffusion. The demic diffusion model stated that farming allowed an increase in the carrying capacity of the regions occupied by farmers and hence allowed rapid demographic growth and spatial expansion toward regions occupied by hunter-gatherers. The latter were then either eliminated by or integrated into the new farming communities that were becoming demographically dominant. The clines observed in the genetic data and their correlation with Neolithic archaeological artifacts (Cavalli-Sforza et al. 1993; Menozzi et al. 1978) or linguistic data (Barbujani and Pilastro 1993) were thus most likely caused by the arrival of the Neolithic farmers in Europe, not by the early arrival of Paleolithic hunter-gatherers (Barbujani et al. 1995). When everything seemed settled in favor of the demic diffusion model, a new and exciting paper was published by Richards et al. (1996). In this study the investigators were claiming that (1) no clear spatial pattern could be seen in mtDNA data and (2) it was possible to define groups of mtDNA haplotypes (haplogroups) and use their age to quantify the relative contributions of Paleolithic and Neolithic people. Based on these rather unusual calculations (see later discussion), they concluded that the contribution of Neolithic farmers to the European gene pool was only of the order of 20%. The fact that this result went against an established consensus made it particularly attractive and thought-provoking. Richards et al.'s (1996) study, however, had a certain number of problems. I only mention this study because it is in many ways the first of a series. Thus it has a historical importance. First of all, when a new method is introduced, it is common practice among population geneticists to demonstrate that it does indeed work as it is supposed to. This rather bizarre tradition has its advantages...