Studies of crystallization of polyethylene (sharp fractions) under hydrostatic pressure revealed that the previously identified mobile hexagonal phase (h) can arise as a metastable transient even in the more customary orthorhombic (o) stability regime. In fact, it was found that crystallization in the h phase controls the crystallization process even there, at least in the portion of the P‐T phase diagram explored so far. Accordingly, crystallization (nucleation and growth) can only take place while in the h phase and ceases upon transformation into the o phase (occurring when the o phase is the stable one). Crystal growth includes both lateral and thickening growth. Thickening growth in particular involves chain refolding, envisaged by sliding diffusion, and can proceed to thicknesses far in excess to chain extension. All growth rates can be defined quantitatively and correlated with morphology. The above findings are raising wider issues such as: Are metastable transients necessary for polymer crystallization under all conditions, or alternatively are there at least two different modes of crystallization with a change‐over unspecifiable by thermodynamics? Beyond this it raises the possibility of a new source for finite uniform lamellar thickness in the form of crystal‐thickening‐induced growth limiting phase inversion, an example for which has already been found in polymer trans‐1,4‐polybutadiene.