IT was first observed by Gross1 that, when the 4358-A. line of mercury scattered by quartz is examined with a 30-step echelon grating, the line is found to be split up into a doublet, each component being shifted from the centre of the line by 0.18 A. He reported later2 that the line is split up into six components. Krishnan and Chandrasekharan3 have recently studied the structure of the 2536.5-A. line of mercury scattered by quartz and filtered through mercury vapour, using a 3-metre quartz spectrograph having a dispersion of 13 cm.−1 per mm. in this region. They have observed a weak doublet, each component of which is shifted from the centre of the line by different amounts for different settings of the crystal. With the direction of incidence along the Y-axis and scattering along the Z-axis of the crystal the shift is 1.90 cm.−1. This shift is too large to be accounted for by the difference in the refractive indices for 4358 A. and 2536 A. The intensity of the displaced components is also only a small fraction of that of the undisplaced line. Hence these results are not in agreement with those observed by Gross.