Measured cold induced vasodilatation (CIVD) temperature and pulse volume changes were used to validate a lumped-parameter heat exchange model of the finger-tip. The simultaneously measured data were extracted from a previous study by Blaisdell (1951).Finger-tip temperatures were estimated using adjusted blood perfusion data as an input. The model was also applied to the inverse problem, namely, to estimate blood perfusion from measured temperatures.A three-segment arterial blood supply temperature profile, composed of an exponential drop followed by a levelling-off and succeeded by an exponential rise, was assumed in order to improve the predictions of the model.Comparison of temperature records showed good conformity, with the estimated values generally under predicting the measured ones. The quality of the predictions of the inverse problem was not as good with local values amounting to less than half the measured ones.In particular the model was shown to retain the basic features of the thermal responses of the cold-stressed extremity, i.e. shape of the curves and the delayed rise of temperature in relation to blood perfusion.