Instantaneous pulsatile velocity profiles were measured in a cylindrical tube using a 20 MHz pulsed Doppler ultrasound system. The spatial resolution of the device was improved with deconvolution using a procedure which employed frequency-domain windowing to suppress high-frequency noise. After deconvolution both the instantaneous velocity profiles and flow waveforms were compared with established theoretical predictions and good agreement was obtained. The average error in a velocity measurement was 7.9 ± 0.9%, and the average error in the flow rate was 4.7 ± 1.7%> The wall shear rate was determined from the gradient of the velocity profile both before and after deconvolution. The results were compared to the theoretical value of wall shear rate to quantify the accuracy of the technique under pulsatile flow conditions. Although the velocity profiles agreed well with theory, the wall shear rate was obtained less accurately. The average error in the mean wall shear rate and peak-to-peak wall shear rate was 28 ± 7% and 11 ± 3%, respectively. However, some of this discrepancy, was attributed to the method of evaluation. In all cases deconvolution significantly improved the accuracy of the wall shear rate determination.