We show that the spacing distributions of low-lying nuclear energy levels recently published by Garrett and collaborators can be represented by a formula previously suggested for the level spacing of systems in transition between order and chaos, equally well if not better than the Brody and Berry-Robnik formulae.