IN QUARTZ OF LIGHT IN THE VISIBLE REGION OF THE SPECTRUM. 
293 
towards the red on increasing the brightness of the lines. Thus sodium light, produced 
by placing a piece of sodium bicarbonate on the grid of a Meeker burner, gave the 
reading 562°‘84, where light from a bead heated by an oxygen-jet in the flame of the 
burner (fig. 2a) gave 562°'80, the light being purified in each case by a D.V. prism on 
the eye-piece of the polarimeter; the mean of these two readings was used in the 
calculation of the optical mass-centre. 
The observations of Zollner (‘ Landolt,’ p. 408) to the effect that the Dg line 
widens symmetrically whilst Dj widens more rapidly towards the red, are likewise in 
accord with the experience gained in the present experiments, where steady values for 
Dg were obtained in the very first series of readings, whilst Dj, although only slightly 
weaker, gave very variable readings and was by far the most troublesome line of the 
whole series to read. 
Whilst there is no great satisfaction, in view of the uncertainties as to its mass- 
centre, in giving figures for the sodium doublet, there is no other way of securing a 
comparison with the observations of other workers. The figures for light of wave¬ 
length 5892’64 are 
Lowry . 
rH Quartz 217270 
IL „ 217286 
Tabulated value 217283. 
Comparison in the form given above with the observations of Soret and Sarasin 
is difficult, as their values for and Dg differ by 0°‘028 per mm. in one case and 
0°‘043 per mm. in the other and are obviously uncertain even in the second decimal; but 
Soret and Guye, working with Soret and Sarasin’s Quartz “ IV ” found the value 
217195, as compared with the values 217205 and 217132 which they quote from 
Soret and Sarasin ; the mean value ma,y be taken as 21718. Landolt has calculated 
the values 21°724 from the readings of VoN Lang ; 21°723 from the readings of Soret 
and Guye; 21°724 from readings given by Gumlich for plates from 5 to 10 mm. in 
thickness, the individual values ranging from 21°717 to 21°731 : and 21°722 from 
the readings of Schonrock for a plate only 5 mm. thick. The general average, 
21°723 is a little lower than the value given by the faulty specimens of quartz rejected 
in the course of the present investigation, for which the value 217255 may be taken 
and indicates that these were rather above than below the general average of quality 
of the specimens used in earlier investigations. 
Some interest attaches to the dispersion-ratios of the defective cylinders of quartz 
which were rejected at the commencement of 1911. The best values for the extreme 
red and violet lines were 
Li . . . 0-647589 instead of 0'647525 (-f 0-000064) 
Hg . . . 1-626753 „ „ 1-626992 (- 0-000239). 
The difierences for other lines were almost all positive towards the red and negative 
towards the violet. The behaviour of these specimens was in very close agreement 
