The purpose of this study was to estimate the time needed to warm an extremity prior to measuring nerve conduction. In 8 normal subjects tibial and sural nerve conduction variables were measured during cooling and warming of the leg in water of 18 degrees C and 36 degrees C, respectively. During cooling, nerve conduction velocity (NCV) decreased and distal motor latency (DML), duration, and area of the compound muscle action potentials (CMAP), and compound nerve action potentials (CNAP) increased. The reverse occurred during warming. During cooling or warming the change in these variables became progressively smaller with time. The time course could therefore be described by an exponential relation, the parameters of which were determined. On the basis of these data, correlations were calculated between the skin temperature at the start of the investigation and the warming time needed to obtain a good estimate of nerve conduction variables at 36 degrees C. The use of correction factors, instead of actual warming, yielded acceptable errors only for NCV and not for the other variables.