A finger cooling test is used to objectify Raynaud's phenomenon and to assess its severity. For this purpose, finger skin temperature (FST) and laser Doppler flux (LDF) are measured on the fingertips before and during cooling of the hand (16 degrees water bath for 5 min) and a subsequent recovery period (20 min). To study reproducibility, this test was performed twice within 3 months in 34 healthy subjects and in 56 Raynaud patients. Three test parameters were used in the analysis: the baseline value, the value after 12 min of recovery, and the mean level during recovery. We determined the limits of agreement (mean differences between the first and the second test +/- 2SD) and the coefficients of variation. No systematic differences between the first and the second test were found. Outside temperature did not influence FST or LDF. FST was shown to have a better reproducibility than LDF. For the baseline value in the total group the coefficient of variation was 3.3% for FST and, rather high, 21.6% for LDF. The limits of agreement for the baseline value in the total group were -4.8 to 4.2 degrees for FST and -25.2 to 22.2 arbitrary units for LDF. In conclusion, the applied cooling test has limited value in individual cases, but can be useful when comparing large groups in pathophysiologic or therapeutic studies.