Experimental studies have demonstrated that demagnetization temperatures for removing thermoviscous magnetizations (TVRMs) in rocks are often higher than predicted by Pullaiah et al.'s time-temperature (t-T) nomogram based on N¿el's single domain (SD) theory. Better agreement has been found for t-T relationships based on Walton's calculations invoking volume distributions. It has been pointed out, however, that Walton's formula applies for t-T conditions for acquisition of magnetization. only, not to demagnetization. We applied numerical methods for assesing the true influence of distributions in SD volume. As it turns out, for reasonable volume distributions, the differences between our results and Pullaiah et al.'s simpler approach are marginal; the differences in demagentization temperature are significant only for moderate-term low-temperature acquisition. Discrepancies between experiments and theory appear because rocks contain predominantly pseudo-single and multi-domain grains to which SD theory does not apply. ¿ American Geophysical Union 1988