Structure optimization, ab initio molecular dynamics (AIMD) simulation of transitions between structures, vibrational self-consistent field (VSCF) calculations of vibrational spectra and infrared ion dip (IRID) experiments have been used to explore the potential energy landscape of isomeric xylose·H 2 O (and D 2 O). The VSCF predictions are in close correspondence with the experimental data but the spectra associated with their two low energy isomers are too similar to permit an unequivocal structural assignment. At cryogenic temperatures several low energy isomers could be ‘frozen in’ but at 300K the AIMD simulations predict rapid transitions between them and in consequence, a highly fluxional system.