No CrossRef data available.
Published online by Cambridge University Press: 25 July 2018
The value obtained for the difference in radial velocity between components of a visual binary that is unresolved on the spectrograph slit appears to depend, sometimes, on whether the measurement is made visually with a microscope or by means of an oscilloscopic setting device. This apparent dependence has been confirmed by measurement of artificially produced doublelined spectrograms, and disappears for line pairs separated by more than 1.5 or 2 times the half-widths of their components. The dependence arises from blending of the two line profiles. There is some evidence that the results obtained from visual measures are affected by the scale of the image being measured. For this reason, oscilloscope measures are probably to be preferred; although their errors are sometimes larger, they seem to be more consistent. Errors arising from this sort of blending are not sufficient to explain measures of relative radial velocity of the components of some visual binaries, that differ widely from the predictions made from the visual orbits.