him to
assign to it a velocity almost identical with that deduced by Roemer
from a totally different method of observation. Subsequently Fizeau,
and quite recently Cornu, employing not planetary or stellar
distances, but simply the breadth of the city of Paris, determined the
velocity of light: while Foucault--a man of the rarest mechanical
genius--solved the problem without quitting his private room. Owing
to an error in the determination of the earth's distance from the sun,
the velocity assigned to light by both Roemer and Bradley is too
great. With a close approximation to accuracy it may be regarded as
186,000 miles a second.

By Roemer's discovery, the notion entertained by Descartes, and
espoused by Hooke, that light is propagated instantly through space,
was overthrown. But the establishment of its motion through stellar
space led to speculations regarding its velocity in transparent
terrestrial substances. The 'index of refraction' of a ray passing
from air into water is 4/3. Newton assumed these numbers to mean that
the velocity of light in water being 4, its velocity in air is 3; and
he deduced the phenomena of refraction from this assumption. Huyghens
took the opposite and truer view. According to this great man, the
velocity of light in water being 3, its velocity in air is 4; but both
in Newton's time and ours the same great principle determined, and
determines, the course of light in all cases. In passing from point to
point, whatever be the media in its path, or however it may be
refracted or reflected, light takes the course which occupies _least
time_. Thus in fig. 4, taking its velocity in air and in water into
account, the light reaches G from I more rapidly by travelling first
to O, and there changing its course, than if it proceeded straight
from I to G. This is readily comprehended, because, in the latter
case, it would pursue a greater distance through the water, which is
the more retarding medium.

Sec. 6. _Descartes' Explanation of the Rainbow_.

Snell's law of refraction is one of the corner-stones of optical
science, and its applications to-day are million-fold. Immediately
after its discovery Descartes applied it to the explanation of the
rainbow. A beam of solar light falling obliquely upon a rain-drop is
refracted on entering the drop. It is in part reflected at the back of
the drop, and on emerging it is again refracted. By these two
refractions, and this single reflection, the ligh