The effect of 2H on 13C longitudinal (T1) and transverse (T2) relaxation parameters was determined for the first time for diastereospecifically deuterium-labeled nucleosides, which are used as the building blocks for non-uniform isotope labeling for the solution NMR structure determination of the large biologically functional oligo-DNA and -RNA (‘NMR window’ approach, ref. 7). It emerged that the T1 and T2 of the deuterated methine carbon in the diastereospecifically deuterium-labeled nucleoside 9 could be used as the correction term to give the monoexponential decay of 13C longitudinal and transverse magnetization of the constituent 1H–13C–2H group. The correlation time derived from this corrected T1 of the methylene carbon corresponds well with the correlation time obtained from deuterium relaxation study. The extreme narrowing limit (ωτc≪1) where dipole–dipole (DD) relaxation of 13C and quadrupole (Q) relaxation of 2H are related by T1DD/T2DD≈1 and T1Q/T2Q≈1 was used to demonstrate the above conclusion. The difference in the observable T1 and T2 in various methylene and methine-type carbons with either fully protonated or diastereospecifically deuterated nucleosides 1–14 allowed the estimation of the contribution of the alternative relaxation pathways other than DD relaxation. It was found by comparison of the T1 relaxation of the quaternary carbon with the methine carbon (13C–2H) or (13C–1H) in compound 2 that the contribution of the intermolecular and intramolecular relaxations of 13C with protons that are two bonds away is larger than DD(13C–2H), and the sum of all these contributions define the T1 of the methine carbon (13C–2H). The observed difference between the experimental T1 and T2 of the methine carbon is attributed to the cross-correlation between DD(13C–2H) and Q(2H) relaxation, which is consistent with recent theoretical predictions. For T2 measurement, the decoupling of deuterium with 0.6–2.5 kHz power during the echo period by WALTZ does not effectively eliminate the DD(13C–2H)–Q(2H) cross-correlation for the methine carbon. The suppression of this DD(13C–2H)–Q(2H) cross-correlation was, however, more effective by applying a 180° deuterium pulse in the middle of the short (0.5 ms) echo period (compare T2 of 3.91s and 0.3s, respectively, at 294 K using these two different decoupling procedures). The comparison of the observed T1 and T2 relaxations of the methylene carbon shows that they are indeed very close. The various contributions of the methine carbon relaxation such as DD(13C–2H), intermolecular and cross-correlation, DD(13C–1H)–Q(2H), to the relaxation of the methylene carbon were ca. 15% in T1 and ca. 25% in T2. © 1998 John Wiley & Sons, Ltd.