Pulsed NMR is a method recommended by international standards for determination of solid fat content (SFC). Fats are thermostated before SFC are determined. Fats are divided into several groups owing to differences in their crystallization processes. Different thermostating procedures have been worked out for each group. According to the ISO 8292 standard on the determination of SFC, fats are divided into four groups: (1) cocoa butter and its equivalents, (2) palm oil, (3) tallow and its fractions, and (4) other fats. According to other standards (IUPAC, AOCS, PORIM), fats are divided into two groups: (1) cocoa butter and its equivalents, and (2) other fats. Therefore, two different procedures are used to thermostate palm oil and tallow. The ISO 8292 standard recommends that these fats are tempered for many hours, whereas other standards recommend only 1 hr cooling at 0°C. In this study, SFC in fats that, according to ISO 8292, belong to four different groups (cocoa butter, palm oil, tallow, and milk fat) were determined. Solid fat contents in samples were determined after a 1 hr cooling and many hr lasting tempering processes. With regard to cocoa butter, palm oil and tallow, ISO procedures specifically designed for these fats were used. Milk fat was tempered in a similar way as cocoa butter. Fats are tempered in order to change them into a stable crystalline form. In case of fats in which such changes occur, i.e. cocoa butter, palm oil and tallow, tempering causes an increase in SFC. In fats not requiring tempering, e.g., milk fat, this step causes a decrease in SFC at a temperature appearing lower as the temperature at which the fat was tempered. Additionally, some standards (IUPAC, PORIM) include a possibility of taking serial measurements. When procedure applied is not suitable for a given fat, i.e. when it is not possible to change the fat into a stable crystalline form, then, a higher SFC is obtained while taking series measurements. This effect is caused by polymorphic transformations occurring during the process of thermostating samples at successively increasing temperatures. If a procedure applied is fitting, both the parallel and serial measurements supply the same results.