Reproduction bottlenecks have been proposed to create a reductive pressure on evolving genomes [1]. Here, we used the aevol model [2] to test the effect of bottlenecks on genome size. AEvol is an artificial evolution model in which organisms can evolve both their phenotypic complexity and structure and their genotypic complexity and structure. A classical experiment with aevol consists in an initialisation phase (during which artificial organisms are created by artificial evolution) and an experimental phase during which the evolutionary dynamics of their genomes can be observed and analysed. In the classical aevol model, the population size is fixed along the whole evolutionary process. Thus, we first modified the aevol model in order to be able to change dynamically the population size. Then, we initialized our experiments by letting 10 populations of 1024 organisms to evolve during 20000 generations. Starting from the 20001st generation, these populations regularly underwent strong bottlenecks (only 8 randomly chosen individuals were allowed to reproduce) followed by an exponential growing phase until the population recovered its original size. We also let the same initial populations evolve without bottlenecks in order to compare both evolutionary dynamics. All experiments were 50000 generations long (from generation 20000 to generation 70000). During these experiments, we first observed that the bottlenecks have a strong effect on the evolutionary dynamics: whereas control populations continued to evolve toward a better adaptation to their environment, bottlenecks constrained the evolution and the organisms’ fitness remained stuck during the 50000 generations of the experiment. As long as genome size is concerned, we observed a clear effect of bottlenecks on genome size but surprisingly, toward a growth of the genomes. While the genomes of control populations slowly shrunk during the experiments (from 2900 bp to 2000 bp), those which underwent bottlenecks grew during the 20000 first generations following the first bottleneck (from 2900 bp to 3800 bp). Then the genomes sizes remained quite constant until the end of the experiment (from generation 40000 to generation 70000). We previously showed that, in aevol, the genome size is be driven by both mutation rates and selection strength [3]. We thus investigated whether the observed effect was due to the difference of selection induced by the bottlenecks. It is known that, in varying population size, the mean selection strength is equivalent to a constant population where the effective population size is equal to the harmonic mean of the population size (76 in our experiment). Using aevol, we verified that stable populations of 76 individuals underwent the same genome growth, thus showing that in our experiments the genome size is governed by the selection strength. We now plan to conduct supplementary experiments in which the populations are submitted to both bottlenecks and deletion bias. We suspect that, depending on the strength of the bottlenecks, the effects of the selection strength and of the deletion bias will interfere, resulting in a complex equilibrium of the genome size that may explain the various situations observed in nature. [1] Nilsson, A.I., Koskiniemi, S., Eriksson, S., Kugelberg, E., Hinton, J.C.D., and Andersson, D. I. (2005). Bacterial genome size reduction by experimental evolution. PNAS, 102(34): 12112-12116 [2] C. Knibbe (2006) Evolution of genome structure by indirect selection of the mutational variability – A computational approach, PhD Thesis, INSA of Lyon, France [3] Knibbe, C., Coulon, A., Mazet, O., Fayard, J.-M., and Beslon, G. (2007) A Long-Term Evolutionary Pressure on the Amount of Noncoding DNA, Mol. Biol. Evol., 24(10): 2344-2353