62 Recent Advances in Stellar Astronomy
a fixed percentage of the parallax, the spectroscopic
method is much superior to the other in accuracy for faint
stars at great distances, where the parallax is smaller than
the errors of observation by the direct method. For the
nearest stars, on the other hand, the trigonometric method
is the more precise. The most serious limitation of the
new method is that, as developed at present, it is avail-
able only for stars of spectrum more advanced than A8.
Some recent work by Lindblad at Mount Wilson, how-
ever, indicates that, by the use of other spectral lines
and somewhat different methods the А-stars may be in-
cluded. As, thanks to Kapteyn, we already know a great
deal about the В-stars, this will fill the last serious gap
in our information.
Two thousand stellar parallaxes have been determined
at Mount Wilson by this method during the past five
years, and the work goes on actively. Shapley at Har-
vard is just beginning investigations of the same sort,
using the unique collection of photographs accumulated by
Pickering’s untiring energy. It is not too much to hope
that, within a very few years more, we shall know the
distances of all the stars of Class F and redder, which
are visible to the naked eye, and perhaps of many fainter
ones. If Lindblad’s methods fulfil their promise, and
Kapteyn’s investigations on the В-stars continue, we may
arrive before long at a very good idea of the distances
and distribution in space of all the stars which can be
seen without a telescope.
The results so far obtained confirm completely the
existence of the giant and dwarf stars, and their charac-
teristic properties, and give us information about the very
brightest stars that we could hardly have hoped to obtain
otherwise. In all the spectral classes there are some stars