This paper gives new evidence that Upper Palaeolithic societies of western Europe would have interpreted stars as patterns or constellations. It also shows how we can scientifically reconstruct the prehistoric mythology. In Eurasia and North America, the stars in the handle of the Big Dipper are often interpreted as hunters and the Dipper itself as an elk or a bear, killed or pursued. This motif is not known on other continents, nor in the Arctic where the Big Dipper is seen better than everywhere else. The link between the Big Dipper and the Cosmic Hunt could only be explained by particular and very ancient historical links between the corresponding traditions. We postulate that the more two myths diverge, the more distant is their genetic relationship, geographically and temporally. We construct a database including the typological variations of the Big Dipper / Cosmic Hunt versions, where the mythological structural features were coded for their presence (1) or absence (0) in each of the target myths. They were selected to provide broad typological coverage, reflecting the known mythological variations of the myth. The whole data matrix contained 44 binary features coded for 18 Amerindian, Asian and European myths. Then we have treated the feature of the studied stories as character traits distributed among taxonomic units (myths). Indeed, there are many analogies between the ways that genes and myths evolved: e.g. both are system of replicators comprising discrete, heritable and highly conservative units (genes and mythemes); they are naturally or socially selected and generally vertically transmitted ; they evolve by a system of descent with modification ; they split into new versions and sometimes go extinct. These connections give hope that the use of phylogenetic methods will succeed and that historical signal can be found. First, the binary-coded mythological feature were computed thanks to a phylogenetic distances algorithm: Bio Neighbor Joining (tree 1), implemented in SplitsTree4, to find the optimal historical tree for the mythological data (Delta score = 0.3239; Q-residual score = 0.06324; 10.000 bootstrap replications). Then, using Mesquite 2.75, we calculate a value for the tree using the same character data matrix. We calculate the parsimony treelength of the tree and matrix. Character matrices were supplied from data files and the tree was rearranged by subtree pruning and regrafting. Tree 2 shows what we obtained (root: Khanty; treelength for character matrix: 74 ; CI: 0.59; RI: 0.71). This tree and the BioNJ tree are convergent. Delta score, CI and RI are indicative of low horizontal transmission and indicate vertical signal in the data. The results show a clear organisation and a progression: (Greece) => Central Asia => Siberia => British Columbia => North-eastern America. This signal is consistent with what we know of the first human migrations. The Big Dipper interpreted as a cosmic hunt must be even older than 15.000 years, coming from Asia, when North America was populated by the migration across the Bering Strait. The orderly and geographically consistent phylogenetic signal shows that phylogenetically analysed mythological pattern can preserve a signal that is consistent with very old human migrations. Phylogenetic approaches have already provided new insights to the origin and distribution of language and culturally transmitted object but it may be the first time that a geographical distribution pattern allows the phylogenetic reconstitution to reach right back the Palaeolithic. As genes and certain words, some myths seem to evolve slowly enough to have time-depths of at least 15000 years, making them good candidate for both deep reconstruction of ancient migrations and prehistoric meanings. Phylogenetic reconstructions using parsimony and Maximum Likelihood with model Mk1 allow to reconstruct ancestral states of the mythems and of the myth itself. The prehistoric tale may be that a man pursues a deer, and that the animal turns alive into the Dipper. The Cosmic Hunt might have been reflected in the rock art of Karelia, Siberia, the Far East and Northern Mongolia, by the authors of the wounded man of Lascaux and in the Greek and Basque mythologies.
