We have compared ultrasound and Harpenden calipers for the measurement of thickness of non-inflamed and anthralin-inflamed skin. In non-inflamed skin ultrasound gave consistently and significantly lower values (0.87 mm +/- 2.3% s.e.) than calipers (1.1 mm +/- 2.7% s.e.) and the ultrasound/caliper ratio was 0.89 +/- 1.3%. In anthralin-inflamed skin the ultrasound/caliper ratio was not significantly different from unity (1.02 +/- 1.3%), but ultrasound was unable to measure inflamed skin thickness in 18% of cases because the echo could not detect the dermis-subcutaneous fat interface. The proportion of unreadable ultrasound results increased linearly with increase in skin thickness and the variance of ultrasound readings increased as inflammatory skin thickness increased; by contrast caliper variance remained constant. Harpenden calipers cannot be used in all subjects or in skin sites in which skin folds cannot be raised, but if the patient and site to be studied can be selected, Harpenden calipers have a greater precision, reproducibility and reliability than ultrasound for measurement of the thickness of inflamed skin.