The point-area method for deconvolution derives a "staircase" input function which, when convolved onto the characteristic function, gives an output function coincidental with the given output data points. The area--area method for deconvolution is shown to be erroneous.