d in Gamma Cassiopeiae by
Keeler in 1889; but even with the giant telescope of Mount Hamilton, the
helium-ray was completely invisible.[1392] It made, nevertheless,
capricious appearances at South Kensington during that autumn, and again
October 21, 1894,[1393] while in September, 1892, Belopolsky could
obtain no trace of it on orthochromatic plates exposed with the 30-inch
Pulkowa refractor.[1394] Still more noteworthy is the circumstance that
the well-known green triplet of magnesium (_b_), recorded as dark by
Keeler in 1889, came out bright on fifty-two spectrographs of the star
taken by Father Sidgreaves during the years 1891-99.[1395] No
fluctuations in the hydrogen-spectrum were betrayed by them; but
subordinate lines of unknown origin showed alternate fading and
vivification.

The spectrum of Beta Lyrae undergoes transitions to some extent
analogous, yet involving a different set of considerations. First
noticed by Von Gothard in 1882,[1396] they were imperfectly made out,
two years later, to be of a cyclical character.[1397] This, however,
could only be effectively determined by photographic means. Beta Lyrae is
a "short-period variable." Its light changes with great regularity from
3.4 to 4.4 magnitude every twelve days and twenty-two hours, during
which time it attains a twofold maximum, with an intervening secondary
minimum. The question, then, is of singular interest, whether the
changes of spectral quality visible in this object correspond to its
changes in visual brightness. A distinct answer in the affirmative was
supplied through Mrs. Fleming's examination of the Harvard plates of the
star's spectrum, upon which, in 1891, she found recorded diverse complex
changes of bright and dark lines obviously connected with the phases of
luminous variation, and obeying, in the long-run, precisely the same
period.[1398] Something more will be said presently as to the import of
this discovery.

Bright hydrogen lines have so far been detected--for the most part
photographically at Harvard College--in about sixty stars, including
Pleione, the surmised lost Pleiad, P Cygni, noted for instability of
light in the seventeenth century, and the extraordinary southern
variable, Eta Carinae. In most of these objects other vivid rays
are associated with those due to hydrogen. A blaze of hydrogen,
moreover, accompanies the recurring outbursts of about one hundred and
fifty "long-period variables," giving banded spectra of the