Studies investigating the effect of rate of temperature change on thermal thresholds have used a variety of different methods and threshold combinations, and many display incomplete reporting of statistical analyses. It has been suggested that C- and Adelta-fibre mediated thresholds differ in their reaction to different rates of temperature change. Ten healthy female volunteers (aged 18-26 years; mean 21 +/- S.D. 2.53) undertook cold sensation (CS), warm sensation (WS), cold pain (CP) and heat pain (HP) threshold determinations on the thenar eminence of the dominant hand. Rates of temperature change of 0.5, 1, 2.5 and 4 degrees C/s were used, with a modified method of limits. Adaptation temperature was 32 degrees C and thermode size 3 cm x 3 cm. Results showed a significant increase in WS, HP and CP thresholds with increased rates of temperature change (all p < 0.001), but no significant change for CS (p = 0.653). These results suggest that thresholds with a C-fibre component (WS, HP and CP) and those that are Adelta-fibre mediated (CS) behave differently. A traditional explanation of measurement artefact alone is insufficient in rationalizing these results, with additional factors potentially involved. Slow rates of temperature change were shown to reduce mean intra-individual differences in recorded threshold values, and also to abolish ceiling effects with HP threshold determinations. Clinically, therefore, using slow rates of temperature change with method of limits has a range of benefits over and above simply minimizing measurement artefact.