5] Its duration was about two seconds, and
the impression produced was that of a complete reversal of the
Fraunhofer spectrum--that is, the substitution of a bright for every
dark line.

Now something of the kind was theoretically necessary to account for the
dusky rays in sunlight which have taught us so much, and have yet much
more to teach us; so that, although surprising from its transitory
splendour, the appearance could not strictly be called "unexpected."
Moreover, its premonitory symptom in the fading out of these rays had
been actually described by Secchi in 1868,[526] and looked for by Young
as the moon covered the sun in August 1869. But with the slit of his
spectroscope placed _normally_ to the sun's limb, the bright lines gave
a flash too thin to catch the eye. In 1870 the position of the slit was
_tangential_--it ran along the shallow bed of incandescent vapours,
instead of cutting across it: hence his success.

The same observation was made at Xerez de la Frontera by Mr. Pye, a
member of Young's party; and, although an exceedingly delicate one, has
since frequently been repeated. The whole Fraunhofer series appeared
bright (omitting other instances) to Maclear, Herschel, and Fyers in
1871, at the beginning or end of totality; to Pogson, at the break-up of
an annual eclipse, June 6, 1872; to Stone at Klipfontein, April 16,
1874, when he saw "the field full of bright lines."[527] But between the
picture presented by the "veritable pluie de lignes brilliantes,"[528]
which descended into M. Trepied's spectroscope for three seconds after
the disappearance of the sun, May 17, 1882, and the familiar one of the
dark-line solar spectrum, certain differences were perceiving, showing
their relation to be not simply that of a positive to a negative
impression.

A "reversing layer," or stratum of mixed vapours, glowing, but at a
lower temperature than that of the actual solar surface, was an integral
part of Kirchhoff's theory of the production of the Fraunhofer lines.
Here it was assumed that the missing rays were stopped, and here also it
was assumed that the missing rays would be seen bright, could they be
isolated from the overpowering splendour of their background. This
isolation is effected by eclipses, with the result--beautifully
confirmatory of theory--of _reversing_, or turning from dark to bright,
the Fraunhofer spectrum. The completeness and precision of the reversal,
however, could not be visually atte