EPR measurements were made at Q-band microwave frequencies on powder samples pressed into pellets and X-irradiated at 4 K. Measurements were made at 4 K after no anneal, then after a 77 K anneal, and then after a 300 K anneal. In poly(dA):poly(T) the free radical distribution is approximately a simple sum of the distributions in the separate homopolymers. In poly(dA-T) the free radical distribution differs from that of poly(dA):poly(T). The clearest difference is that in poly-(dA):poly(T) the concentration of one-electron-reduced thymine (Tre.) is reduced relative to the total radical concentration. On warming the thymine-containing samples from 77 K to room temperature, the Tox. radical disappears and the (T-H5'). radical appears. Also, the Tre. radical disappears and the (T + H6). radical appears. There are three main conclusions. First, little or no transfer of free radicals between strands is needed to explain the data. Second, when A and T are interstacked, either the Tre. radical is less stable against recombination than other radical products or radical transfer occurs to an adjacent adenine. Third, in the poly- and oligonucleotides, the Tox. radical is a likely precursor to the (T-H5'). radical.