the position of the spectrum on the first
screen could be shifted sufficiently to cause light of any desired
colour to pass through. Some of this light also passed through the
second hole, and thus he obtained a narrow beam of practically
homogeneous light in a fixed direction (the line joining the apertures
in the two screens). Operating on this beam with a second prism, he
found that the homogeneous light was not dispersed, and also that it was
more refracted the nearer the point from which it was taken approached
to the violet end of the spectrum RV. This confirmed his previous
conclusion that the rays increase in refrangibility from red to violet.

[Illustration: FIG. 2.]

Newton also made use of the method of crossed prisms, which has been
found of great use in studying dispersion. The prism P (fig. 3) refracts
upwards, while the prism Q, which has its refracting edge perpendicular
to that of P, refracts towards the right. The combined effect of the two
is to produce a spectrum sloping up from left to right. The spectrum
will be straight if the two prisms are similar in dispersive property,
but if one of them is constructed of a material which possesses any
peculiarity in this respect it will be revealed by the curvature of the
spectrum.

[Illustration: FIG. 3.--Method of Crossed Prisms.]

The coloured borders seen in the images produced by simple lenses are
due to dispersion. The explanation of the colours of the rainbow, which
are also due to dispersion, was given by Newton, although it was known
previously to be due to refraction in the drops of rain (see RAINBOW).

According to the wave-theory of light, refraction (q.v.) is due to a
change of velocity when light passes from one medium to another. The
phenomenon of dispersion shows that in dispersive media the velocity is
different for lights of different wave-lengths. In free space, light of
all wave-lengths is propagated with the same velocity, as is shown by
the fact that stars, when occulted by the moon or planets, preserve
their white colour up to the last moment of disappearance, which would
not be the case if one colour reached the eye later than another. The
absence of colour changes in variable stars or in the appearance of new
stars is further evidence of the same fact. All material media, however,
are more or less dispersive. In air and other gases, at ordinary
pressures, the dispersion is very small, because the refractivity is
small. The dispe