From the infrared spectra of the vapor measured between 4000 and 200 cm-', 2-cyanoethanol was found to exist as garrche and trnns isomers in roughly equal proportions around 100 "C. The gartclle form is the more stable of the two, as in 2-chloro-and 2-bromo-ethanol, but the enthalpy difference is less than half as large; at most 0.7 kcal mole-'. This indicates some weak (about 1 kcal mole-') intramolecular hydrogen bonding between the O H group and the 71 electrons of the C=N bond.In the solid, two distinct crystalline phases have been observed. Crystal 11, consisting of gauche molecules only, is normally obtained; for instance by annealing the vitreous solid froni low teniperature deposition of the vapor. Crystal I, which contains the two ison~ers, may be obtained irreversibly froni the melt of crystal I1 under controlled conditions.On the other hand there was no indication in the spectra of a cyclic tautomer as suggested recently.Les spectres infrarouges de la vapeur pris entre 60 et 140 "C ont confirm6 la presence dans le 2-cyanoethanol (CN-CH2-CH2-OH) de deux isonieres de rotation, le garrche et le trmis, en proportions a peu pres egales. L'isomere garrche est le plus stable des deux, comme dans le 2-chloro-et le 2-bromoethanol; cependant la difference d'enthalpie est moins de la moiti6 de ceux-ci, soit 0.7 kcal mole-' tout au plus. Ceci doit Ctre attribui a une liaison hydrogene intramoleculaire faible (environ 1 kcal mole-') entre le groupe OH et les electrons n de la liaison C k N .Tout commele 2-chloroethanol, le 2-cyanoCthano1 peut exister sous deux phases cristallines distinctes. La phase normale (11) faite de molCcules galrcl~es seulement, peut se transformer par fusion menagte en line phase metastable (I) qui contient les deux isomeres.Par ailleurs on ne trouve dans les spectres aucune indication d'une forme tautomere cyclique telle que proposee rCceniment .