have now some difficult work before us. Hitherto we have been
delighted by objects which addressed themselves as much to our
aesthetic taste as to our scientific faculty; we have ridden
pleasantly to the base of the final cone of Etna, and must now
dismount and march through ashes and lava, if we would enjoy the
prospect from the summit. Our problem is to connect the dark lines of
Fraunhofer with the bright ones of the metals. The white beam of the
lamp is refracted in passing through our two prisms, but its different
components are refracted in different degrees, and thus its colours
are drawn apart.

Now the colour depends solely upon the rate of oscillation of the
atoms of the luminous body; red light being produced by one rate, blue
light by a much quicker rate, and the colours between red and blue by
the intermediate rates. The solid incandescent coal-points give us a
continuous spectrum; or in other words they emit rays of all possible
periods between the two extremes of the spectrum. Colour, as many of
you know, is to light what _pitch_ is to sound. When a violin-player
presses his finger on a string he makes it shorter and tighter, and
thus, causing it to vibrate more speedily, heightens the pitch.
Imagine such a player to move his fingers slowly along the string,
shortening it gradually as he draws his bow, the note would rise in
pitch by a regular gradation; there would be no gap intervening
between note and note. Here we have the analogue to the continuous
spectrum, whose colours insensibly blend together without gap or
interruption, from the red of the lowest pitch to the violet of the
highest. But suppose the player, instead of gradually shortening his
string, to press his finger on a certain point, and to sound the
corresponding note; then to pass on to another point more or less
distant, and sound its note; then to another, and so on, thus sounding
particular notes separated from each other by gaps which correspond to
the intervals of the string passed over; we should then have the exact
analogue of a spectrum composed of separate bright bands with
intervals of darkness between them. But this, though a perfectly true
and intelligible analogy, is not sufficient for our purpose; we must
look with the mind's eye at the oscillating atoms of the volatilised
metal.

Figure these atoms as connected together by springs of a certain
tension, which, if the atoms are squeezed together, push them again
asu