Finger systolic-pressure (FSP) in response to local cooling has been evaluated as a diagnostic test in 28 females with primary Raynaud's phenomenon (PRP). The mean FSP (%) at 15 degrees C and 10 degrees C was significantly lower in the patients than in 28 age- and sex-matched controls but there was a large overlap in the individual responses. Diagnostic sensitivity of FSP (%) at 10 degrees C was only 57%. Increasing the sensitivity of the test lowered the specificity as the number of false positives rose. This also reduced the predictive value for a positive test. As a diagnostic test of PRP in individuals, the measurement of FSP appears to be of limited value, except in severe cases when complete vessel closure often occurs. Measurement of FSP will classify groups correctly and this objective physiological measure of Raynaud's phenomenon may have a place in group comparisons.