the
dark space falls a portion of the light of the sodium flame.

[Illustration: Fig. 56.]

I have already referred to the experiment of Foucault; but other
workers also had been engaged on the borders of this subject before it
was taken up by Bunsen and Kirchhoff. With some modification I have on
a former occasion used the following words regarding the precursors of
the discovery of spectrum analysis, and solar chemistry:--'Mr. Talbot
had observed the bright lines in the spectra of coloured flames, and
both he and Sir John Herschel pointed out the possibility of making
prismatic analysis a chemical test of exceeding delicacy, though not
of entire certainty. More than a quarter of a century ago Dr. Miller
gave drawings and descriptions of the spectra of various coloured
flames. Wheatstone, with his accustomed acuteness, analyzed the light
of the electric spark, and proved that the metals between which the
spark passed determined the bright bands in its spectrum. In an
investigation described by Kirchhoff as "classical," Swan had shown
that 1/2,500,000 of a grain of sodium in a Bunsen's flame could be
detected by its spectrum. He also proved the constancy of the bright
lines in the spectra of hydrocarbon flames. Masson published a prize
essay on the bands of the induction spark; while Van der Willigen, and
more recently Pluecker, have also given us beautiful drawings of
spectra obtained from the same source.

'But none of these distinguished men betrayed the least knowledge of
the connexion between the bright bands of the metals and the dark
lines of the solar spectrum; nor could spectrum analysis be said to be
placed upon anything like a safe foundation prior to the researches of
Bunsen and Kirchhoff. The man who, in a published paper, came nearest
to the philosophy of the subject was Angstroem. In that paper,
translated by myself, and published in the "Philosophical Magazine"
for 1855, he indicates that the rays which a body absorbs are
precisely those which, when luminous, it can emit. In another place,
he speaks of one of his spectra giving the general impression of the
_reversal_ of the solar spectrum. But his memoir, philosophical as it
is, is distinctly marked by the uncertainty of his time. Foucault,
Thomson, and Balfour Stewart have all been near the discovery, while,
as already stated, it was almost hit by the acute but unpublished
conjecture of Stokes.'

Mentally, as well as physically, every year of the