Archived data for Broquet T, Viard F and Yearsley J. M., Genetic drift and collective dispersal can result in chaotic genetic patchiness", Evolution Thomas Broquet, sept. 2012. Archived text files are the parameter files used with simulation software Nemo to create the raw genetic data used in the manuscript. The reference for Nemo is: Guillaume, F., and J. Rougemont. 2006. Nemo: an evolutionary and population genetics programming framework. Bioinformatics 22:2556-2557. Simulation scenarios are named A1, A2, .... Default parameters are indicated in the main manuscript (table 2). Simulation scenarios are as follows: A1 to A4 : extinction rate (e) varies from 0 to 0.9 B1 to B5 : collective dispersal (psi) varies from 0.1 to 0.9 C1 to C4 : extinction rate and collective dispersal vary together These simulations were used for figures 4-6 and S2-S4 G1 to G3 : deme size N set to 25, 50 or 100; psi=0.5; e=0 H1 to H3 : N set to 25, 50 or 100; psi=0.5; e=0 I1 to I3 : N set to 25, 50 or 100; psi=0.5, e=0 These simulations were used for figure 7. J1 to J4 : migration rate (m) set to 1, 0.99, 0.9, or 0.5 These simulations were used for figure S5 Each simulation was run for a burnin period of 15000 generations (files chaosb_A1.ini, chaosb_A2.ini, etc.), saving data every 200 generations. Then each simulation was run for another 20 generations (starting from the state previously stored at the end of a "burnin" simulation), saving useful data every generation. Only these data were used in the analyses. For these final simulations the parameter files are named chaosA1.ini, chaosA2.ini, etc. The simulations that were used to look at global statistics (e.g. A1 to C4) were run using 100 replicates. Some simulations used to investigate only spatial or temporal pairwise statistics were run for one replicate only (because using 100 pairs from one replicate or from more replicates makes no difference, see section "simulation analyses and results" in the main text). Contact : thomas.broquet@sb-roscoff.fr