The thermal-diffusion factor, αT, for the neon–xenon system has been measured from 328°–873°K at seven compositions ranging from 5%–95% neon. These measurements were made with a 20-tube trennschaukel where the pumping rate, the quantity of gas pumped, the duration of the experiments, and the temperature gradient could be varied over a wide range. This flexibility allowed the operating conditions to be chosen so that the theoretical corrections for trennschaukel operation were small and therefore could be applied to the measured separations with confidence. The present corrected data have been compared with theoretical calculations based on the Lennard-Jones (12–6) potential and the modified Buckingham (exp-6) potential. Both the experimental data and the theory indicate that the reciprocal of αT is a linear function of the composition. The present data agree well with the results of Grew and of Heymann and Kistemaker, but do not agree with that of Saxena and co-workers nor with the results of Atkins, Bastick, and Ibbs. The mutual diffusion coefficient for neon–xenon, which was derived from the measured composition dependence of αT, agrees well with experimental diffusion coefficients reported in the literature and also with values calculated from experimental viscosity and thermal conductivity data. Although the values of αT are in disagreement with the present work, the composition dependence of αT reported by Saxena, Nain, and Saxena also appears to be consistent with the values of the diffusion coefficient; however, that reported by Atkins et al. and Mathur and Saxena are not.