The photographic subtraction formula expressing the 'Doppler signal' in terms of line-ofsight velocity is rederived and the conditions governing its validity are carefully stated. By appealing to the observed profiles of individual bright and dark mottles it is shown that, in the case of Hc~ observations of the quiet chromosphere, two essential conditions are violated.
An analysis based on Beckers' (1964) theoretical mottle profiles shows that, assuming the correctness of his model, the photographic subtraction method correctly maps regions of zero velocity and, provided the original photographs are taken sufficiently far from the centre of Hcq gives the correct signs of any line-of-sight velocities present. However, at Hc~ β’ 0.25 A the method gives inconsistent signs. Moreover, at all wavelengths in the line the magnitude of the Doppler signal depends not only on velocity but also on the source function S, optical thickness to, and line broadening parameter A%0. Accordingly, there is no one-to-one correspondence between velocity and the Doppler signal since S, in particular, shows considerable variation from place to place over the quiet chromosphere.
An alternative procedure which obviates these difficulties is suggested. It is also based on Beckers ~ model and yields a calibrated measure of the velocity in terms of A20 (assumed constant) which is independent of spatial variations in S and to.