A chemist friend of mine did his thesis on lipid vesicles, and I remember my mind being blown when he told me these are modelled as a liquid on the 2D plane of the membrane, but as a solid on the 1D orthogonal direction because the energy to swap two lipid molecules side by side is incredibly low (because it makes barely any difference), while the energy to swap them orthogonally to the membrane is much larger (because they would point in the wrong direction).