Firstly, we see that the analytical predictions compare well to integrating the full stochastic differential equations, validating our simplifying assumptions. Secondly, we see our coarse-grained continuous model predicts that there is a large population size effect for the probability of an incompatibility, where the characteristic time for incompatibilities to arise becomes much shorter as the population size decreases. In addition, it predicts that the dynamics of PI(t ) become insensitive to differences in population size both for small population sizes (κFNe?1/?κFNe?1/?) and large population sizes (κFNe?1/?κFNe?1/?).