Title: On the secular evolution of the ratio between gas and dust radii in protoplanetary discs

Abstract: A key problem in protoplanetary disc evolution is understanding the efficiency of dust radial drift. This process makes the observed dust disc sizes shrink on relatively short timescales, implying that discs started much larger than what we see now. In this paper we use an independent constraint, the gas radius (as probed by CO rotational emission), to test disc evolution models. In particular, we consider the ratio between the dust and gas radius, $R_{\rm CO}/R_{\rm dust}$. We model the time evolution of protoplanetary discs under the influence of viscous evolution, grain growth, and radial drift. Then, using the radiative transfer code RADMC with approximate chemistry, we compute the dust and gas radii of the models and investigate how $R_{\rm CO}/R_{\rm dust}$ evolves. Our main finding is that, for a broad range of values of disc mass, initial radius, and viscosity, $R_{\rm CO}/R_{\rm dust}$ becomes large (>5) after only a short time (<1 Myr) due to radial drift. This is at odds with measurements in young star forming regions such as Lupus, which find much smaller values, implying that dust radial drift is too efficient in these models. Substructures, commonly invoked to stop radial drift in large, bright discs, must then be present, although currently unresolved, in most discs.