REFRACTIVE INDICES. 87 
The center of the visible spectrum of ordinary daylight is in the yellowish- 
green and if care be taken in these experiments to obtain the same intensity 
of light in the red band as in the blue band, by lowering or raising the 
refractive index of the liquid, it is possible to determine the refractive index 
of the mineral plate for the color about 555 nn with a probable error of 2 or 
3 in the third decimal place. 
If oblique illumination be used the phenomena are even more distinct, 
as the colored red and blue fringes appear on opposite sides of the grain. 
From Fig. 53, a, b, it is evident that in case both liquid and mineral grain 
have the same refractive index for yellow, the red rays are directed toward 
the right side of the grain (the mineral having the higher refractive index 
for the red rays), while the blue rays appear on the left side of the grain.* 
Greater accuracy than =*= 0.002 can not be obtained by methods based on 
relative dispersions, because of certain disturbing factors which it is difficult 
to control satisfactorily. The point of average intensity in the visible day- 
light spectrum is not constant, but shifts its position noticeably on different 
days and under different conditions of weather.f If, therefore, the refrac- 
tive index of the grain be determined by matching the intensities of the 
red and blue bands on the opposite sides of an immersed grain, the value 
obtained will be slightly different on different days. Still another factor is 
the sensitiveness of the eye, which varies with the intensity of illumination, 
the maximum being at about 510 ;uju for weak illumination and increasing 
toward the yellow up to at least 540 /zju.J If th observations be made, 
therefore, with raised condenser and brightly lighted field, the sensitiveness 
of the eye will not be the same as it is for a weakly lighted field (condenser 
lowered and aperture decreased) and slightly different results would be 
obtained under the different conditions. Several of these factors are sub- 
jective in character and not capable of precise measurement by simple 
methods, and prescribe, therefore, a practical limit to the accuracy of the 
determination by such relative dispersion methods. The center of the 
visible spectrum can be shifted by the use of suitable ray filters which 
absorb part of the light from either the blue or red end of the spectrum, as 
was first suggested by Maschke, and the refractive index for a different 
part of the spectrum thus ascertained with about the same probable error, 
0.002. The chromatic errors of the lens system are a third factor which 
exerts a disturbing influence on the color phenomena observed in oblique 
illumination. 
For more accurate work, the color phenomena with their subjective ele- 
ments can not be used and recourse must be had to strong monochromatic 
light. Having once determined the approximate refractive index of a min- 
eral grain by one of the relative dispersion methods, the more exact measure- 
ments are made in in tense monochromatic light either from a monochromatic 
illuminator or Geissler tube or monochromatic Bunsen flame from sodium, 
lithium, thallium, or other suitable salt. Either the Becke line method or 
*It should be noted that, in case the oblique rays are obtained by shading half the condenser with the 
forefinger, the phenomena can be reversed by raising the condenser beyond the point where the finger edge 
is focussed on the object. 
tE. L. Nichols, Phys. Rev.. 26, 497. 1908; 28, 122. 1909: Trans. 111. Soc.. 3, 301. 
JA detailed consideration of the subject of the Luminous Equivalent of Radiation has recently been gives 
by P. G. Nutting. Bull. U. S. Bureau of Standards. 5, 261-308, 1008. 
JWied. Ann. Phys. Chcm., II, 772-734, 1880; ray filters for the same purpose have also been used by 
Dr. Merwin of the Geophysical Laboratory. 
