The Light of the Stars 49
with lower dispersion. Years of study at Harvard, mainly
by E. C. Pickering, Mrs. Fleming and Miss Cannon, have
culminated in the classification of the spectra of more than
200,000 stars.
The familiar, but nevertheless amazing, result of these
researches is that all but a minute fraction of these can
be placed in one or another of six spectral classes, which,
during the progress of the work, came, by the law of
survival of the fittest in nomenclature, to be designated
by the arbitrary letters B, A, F, G, K, and M. What
is still more noteworthy is that these six classes grade
into one another imperceptibly, so that they form parts
of a single sequence (in the order named above). The
transition from any type to the next always takes place
through the same intermediate stages (illustrated, of
course, by different stars) so that we may adopt the
familiar decimal classification, and call a star half way
between B and A, for example, B5A, or simply B5.
This linear sequence of types finds room for more
than 99 per cent, of the stars. The exceptions can al-
most all be placed in three additional classes—denoted
by the letters O, R and N—of which the first evidently
belongs at the head of the list, before B, and is con-
nected with it by intergrades, while the other two form
a sort of side-chain branching from the main sequence
near K.
Spectra near the head of this sequence show mainly
the lines of the permanent gases—hydrogen, helium, oxy-
gen and nitrogen. As we pass on from B to A the
helium lines disappear, and lines of the metals come in
—at first those lines which show in the laboratory only,
or chiefly, in the spectrum of the electric spark. Farther
on, in Class F, the arc lines of the metals appear, and in