find not only that the order of colours is reversed, but also that the
same colours do not occupy corresponding lengths on the two spectra,
the blue and violet being much more extended in the refraction
spectrum. The refraction spectra for different media also differ
amongst themselves. This shows that the connexion between the
refrangibility of light and its wave-length does not obey any simple
law, but depends on the nature of the refracting medium. This property
is referred to as the "irrationality of dispersion." In a diffraction
spectrum the diffraction is proportional to the wave-length, and the
spectrum is said to be "normal." If the increase of the angle of
refraction were proportional to the diminution of wave-length for a
prism of any material, the resulting spectrum would also be normal.
This, however, is not the case with ordinary refracting media, the
refrangibility generally increasing more and more rapidly as the
wave-length diminishes.

The irrationality of dispersion is well illustrated by C.
Christiansen's experiments on the dispersive properties of white
powders. If the powder of a transparent substance is immersed in a
liquid of the same refractive index, the mixture becomes transparent
and a measurement of the refractive index of the liquid gives the
refractivity of the powder. Christiansen found, in an investigation of
this kind, that the refractivity of the liquid could only be got to
match that of the powder for mono-chromatic light, and that, if white
light were used, brilliant colour effects were obtained, which varied
in a remarkable manner when small changes occurred in the refractive
index of the liquid. These effects are due to the difference in
dispersive power of the powder and the liquid. If the refractive index
is, for instance, the same for both in the case of green light, and a
source of white light is viewed through the mixture, the green
component will be completely transmitted, while the other colours are
more or less scattered by multiple reflections and refractions at the
surfaces of the powdered substance. Very striking colour changes are
observed, according to R. W. Wood, when white light is transmitted
through a paste made of powdered quartz and a mixture of carbon
bisulphide with benzol having the same refractive index as the quartz
for yellow light. In this case small temperature changes alter the
refrac