The martensitic transformation of a Cu–Zn–Al single crystal was measured in a differential scanning calorimeter at different heating rates. The curves were analyzed in order to detect the initial and final temperatures of the transition. The results were analyzed using both, the method suggested by Benke et al. and the conventional method of extrapolating the temperatures to zero rate. They were also compared with those determined in the same calorimetric sample using the electrical resistance measurement. It is concluded that the transition happens in a range higher than that detected by calorimetry and that the inflexion point is not a useful point for a proper temperature determination.