The kinetics of the incorporation of tritium into the C-8 positions of purine units in nucleic acids has been studied. The polymers investigated include poly(A), poly(A): poly (U) duplex, a double-stranded viral RNA, tRNA, and DNA. In the random coil state, the kinetics of incorporation of tritium into the purine sites of the polymers are identical with those for the corresponding purine mononucleotides. When the nucleic acids are in their native conformations, however, the purine labeling rates are reduced below that expected for the free mononucleotides. The magnitude of the effect is remarkably dependent upon the particular nucleic acid. For example, at 37 degrees C the purines in double-stranded DNA label at a rate two- to threefold slower than the corresponding mononucleotides, but in a double-stranded viral RNA, a 30- to 40-fold effect is found. The data suggest a strong influence of microscopic helix structure on the rate of tritium incorporation. First-order rate constants for the exchange of tritium into specific purine sites in yeast tRNAPhe were also determined. This was done by partially labeling the nucleic acid in tritiated water, and subsequently removing free and loosely bound tritium. Under conditions where exchange-out does not occur, the nucleic acid was digested with specific nucleases; chromatographic separation then enabled specific activities of purines from specific sites to be obtained. The rate constants for these sites show a large variation. They are markedly reduced for those residues occurring in cloverleaf helical sections and, in certain cases, for those known from crystallographic data to be involved in tertiary interactions. As examples of bases that can participate in tertiary interactions, the crystal structures show A14 and G15 in special base-pairing arrangements. Both purines (A14 and G15) occur in single-stranded sections of the cloverleaf; both show markedly reduced C-8 hydrogen-exchange rates. On the other hand, rate constants for bases and regions known to be on the outside of the moleculesuch as the anticodon loop and the 3' terminusāre perturbed the least. In one instance, a base in the dihydrouridine loop believed to be involved in tertiary interactions, according to crystallographic studies, incorporates tritium as if it were relatively unperburbed by the tRNA structure. The structural interactions of this base may be partially or completely broken at 37 degrees C in solution.