Many crustal magmatic reservoirs are fundamentally powered by basalt injection at their base. Asides from transfer of silicate liquid and crystals and accompanying heat, basaltic magma also provides large amounts of fluids to the overlying magma. This work explores how water activity and temperature, hence degree of crystallisation , of magmatic reservoirs are affected by such a mechanism at various levels in the crust. By using recent experimental phase equilibria , thermodynamic relationships between gas and silicate melts , and heat balance, it is shown that, depending on the level of magma storage, diffusive exchange during bubble uprise and stalling may produce either crystallisation or melting of magmas. Long term fluxing of felsic to intermediate magma bodies stored in upper crust by mafic volatiles will generally lead to their near isothermal solidification. Conversely, for bodies stagnating in the mid to deep crust, such a process almost inevitably enhances melting, driving or maintaining magmas beyond the threshold of mobility needed for upward material transfer, unless the percolating fluid is very CO 2 -rich. Compilation of basaltic melt inclusion data gathered in arc, hot-spot and ridge settings, shows that the two last categories coexist with CO 2 -rich fluids at high pressures (XH 2 O fluid < 0.1), which will almost always enhance crystallisation. In contrast arc basalts record a wide and continuous range of fluid compositions, from dry to almost H 2 O-saturated conditions, which may either favour (low pressure) or inhibit (high pressure) the crystallisation of felsic reservoirs which they underplate. Crustal growth may thus be in part limited by the difficulty of crystallising deep-seated magma bodies, in particular in arc settings. This pressure controlled effect is related to the contrasted solubilities of H 2 O and CO 2 in silicate melts, and to the much stronger non-ideal behaviour of CO 2 relative to H 2 O as pressure increases. Asides from tectonic and density inversion processes, a thick crust may fundamentally reflect the fact that basalt underplating in the lower crust has proceeded at a rate sufficiently slow so as to prevent remelting of earlier intrusions (or melting of lower crust lithologies) or that the outcoming fluid was CO 2 -rich, or both. Application to other terrestrial planets is hindered by the paucity of data regarding crust thickness and composition, and it can be only conjectured that the thin crust of large planets (Earth, Venus) reflects in part the operation of subduction process during their evolution, while the comparatively thicker crust inferred for smaller bodies (e.g. Mars, the Moon, Vesta) primarily reflects processes related to a primordial crust. • Long term fluxing of felsic magma bodies in upper crust by mafic volatiles generally leads to their solidification. • Conversely, in the deep crust, such a process almost inevitably enhances melting, unless the mafic fluid is very CO2-rich. • Crustal growth may thus be limited by the difficulty of crystallising deep-seated magma bodies, in particular in arc settings. • A relatively thin crust on large planets may reflect the operation of subduction