Measurements of graphite have been studied for thermal white cast iron samples, where data from thermal analysis has been compared to graphite contents measured in light optical microscope. Data from the thermal analysis gives random outcome and is therefore not suitable as a measurement of graphite content in white cast iron. The reason for this is the low amount of graphite in white cast iron. Furthermore, the sampling for the thermal analysis must be performed at comparable temperatures in order to provide usable data.Sampling from thermal analysis for nodular cast iron has been compared to samples taken from the rollpin on finished rolls. The result indicates that the thermal analysis samples have a lower amount of graphite compared to the finished rolls. The reason for this is that the thermal analysis samples cool faster than the finished rolls, which are larger and demand longer cooling time.