In order to calculate the mean transit time of tissue, such as brain, from dynamic computed tomography performed after a bolus injection of intravenous contrast material, the time dependence of the input of contrast material to the tissue must be "deconvolved" from the observed time course of the tissue contrast enhancement. If the approximate shape of the curve of the response of the tissue to an instantaneous injection of contrast material is assumed, the width of this curve that gives the best fit to the observed tissue response can be used to find a value for the tissue mean transit time. Applying this technique to dynamic CT scans of two normal volunteers yielded values comparable to those in the literature by other techniques. The method has the advantages of being simple to implement, relatively insensitive to noise and the details of the assumed curve shape, and not requiring any curve fitting to correct for recirculation.