Conservation of population size is required for self-organized criticality in evolution models

We study models of biological evolution and investigate a key factor to yield self-organized criticality (SOC). The Bak-Sneppen (BS) model is the most basic model that shows an SOC state, which is developed based on minimal and plausible assumptions of Darwinian competition. Another class of models, which have population dynamics and simple rules for species migrations, has also been studied. It turns out that they do not show an SOC state although the assumptions made in these models are similar to those in the BS model. To clarify the origin of these differences and to identify a key ingredient of SOC, we study models that bridge the BS model and the dynamical graph model, which is a representative of the population dynamics models. From a comparative study of the models, we find that SOC is found when the fluctuations of the number of species N are suppressed, while it shows off-critical states when N changes according to its evolutionary dynamics. This indicates that the assumption of the fixed system size in the BS model plays a pivotal role to drive the system into an SOC state, and casts doubt on its applicability to actual evolutionary dynamics.