The second preprint from our experimental-theory collaboration between the Aulehla and Erzberger groups at EMBL is now online. Experiments by Anniek Stokkermans and Michael Zhao showed how primary cells from mouse embryos can reform spatial structures after randomization. These structures scale with the overall system size for small systems, but this breaks down for larger systems. We develop a mathematical theory based on nonlocal advection-diffusion equations and obtained parameter estimates for all model parameters from imaging data. We suggest that the scaling we observe is the result of a time-size trade-off, relevant for embryonic development. Find the preprint on bioRxiv.