We examined the temperature dependence of relaxation times in proton components of fatty acids in various samples in vitro at 11 tesla as a standard calibration data for quantitative temperature imaging of fat. The spin-lattice relaxation time, T 1 , of both the methylene (CH 2 ) chain and terminal methyl (CH 3 ) was linearly related to temperature (rÀ0.98, Pº0.001) in samples of animal fat. The temperature coe‹cients for the 2 primary proton components diŠered signiˆcantly; in 5 bovine fat samples, the coe‹cient at 309 C was 1.79 ±0.07 (z/9 C) for methylene and 2.98±0.38 (z/9 C) for methyl. Numerical simulations based on such a diŠerence demonstrated the possibility of considerable error from inconsistent ratios in fatty acid components when calibrating and estimating temperature. The error reached 3.39 C per 159 C in temperature elevation when we used a pure CH 2 signal for calibration and observed the signal with 18z of CH 3 to estimate temperature. Theseˆnd-ings suggested that separating the fatty acid components would signiˆcantly improve accuracy in quantitative thermometry for fat. Use of the T 1 of CH 2 seems promising in terms of reliability and reproducibility in measuring temperature of fat.