The specific heat ${C}_{p}$ of thulium metal has been measured in a ${\mathrm{He}}^{3}$ cryostat. Between 0.38 and 3.9\ifmmode^\circ\else\textdegree\fi{}K ${C}_{p}=2.839{T}^{3}+17.94T+23.43{T}^{\ensuremath{-}2}\ensuremath{-}1.79{T}^{\ensuremath{-}3}\ensuremath{-}0.066{T}^{\ensuremath{-}4}$ (in mJ/mole \ifmmode^\circ\else\textdegree\fi{}K). The last three terms represent the nuclear specific heat ${C}_{N}$. On the basis of earlier estimates, we put ${C}_{L}=0.243{T}^{3}$ and ${C}_{E}=10.5T$ for the lattice and electronic specific heats, respectively. According to the simple spin-wave theory, the magnetic specific heat ${C}_{M}$ is proportional to ${T}^{3}$ for a ferrimagnetic metal; experimentally one finds ${C}_{M}=6.2{T}^{\frac{5}{2}}$ for thulium, which has a rather complicated ferrimagnetic structure. Further, there seems to be no evidence in ${C}_{M}$ for an exponential factor, to be expected because of magnetic anisotropy. All conclusions on ${C}_{M}$ are tentative, however, until data at temperatures between 4 and 20\ifmmode^\circ\else\textdegree\fi{}K become available. ${C}_{N}$ does not fit to the simple picture as given by Bleaney either. Since $I=\frac{1}{2}$ for the only stable thulium isotope ${\mathrm{Tm}}^{169}$, quadrupole interactions are zero and there are only two nuclear energy levels, their separation being determined by the magnetic hyperfine constant ${a}^{\ensuremath{'}}$. This would give a nuclear specific heat with even powers of $T$ only, with ${a}^{\ensuremath{'}}$ determining the values of the coefficients. The observed ${C}_{N}$ cannot be fitted into an equation of this type which indicates that other interactions, probably nuclear exchange interactions, are present. Formally, the experimental situation may be expressed by writing ${a}^{\ensuremath{'}}={a}_{0}\ensuremath{-}\frac{b}{T}$, instead of treating ${a}^{\ensuremath{'}}$ as a constant. Our results are in good agreement with recent M\"ossbauer data by Kalvius et al. who found 22.9 for the coefficient of the ${T}^{\ensuremath{-}2}$ term.