Using the spectral indices as an
expression for the spectral types we find that the mean spectral index
of these stars is +1.1 corresponding to the spectral type F1.

31. _Stars with the greatest proper motion._ In table 3 I have collected
the stars having a proper motion greater than 3" per year. The
designations are the same as in the preceding table, except that the
names of the stars are here taken from different catalogues.

In the astronomical literature of the last century we find the star 1830
Groombridge designed as that which possesses the greatest known proper
motion. It is now distanced by two other stars C. P. D. 5h.243
discovered in the year 1897 by KAPTEYN and INNES on the plates taken for
the Cape Photographic Durchmusterung, and BARNARD's star in Ophiuchus,
discovered 1916. The last-mentioned star, which possesses the greatest
proper motion now known, is very faint, being only of the 10th
magnitude, and lies at a distance of 0.40 sir. from our sun and is
hence, as will be found from table 5 the third nearest star for which we
know the distance. Its linear velocity is also very great, as we find
from column 10, and amounts to 19 sir./st. (= 90 km./sec.) in the
direction towards the sun. The absolute magnitude of this star is 11m.7
and it is, with the exception of one other, the very faintest star now
known. Its spectral type is Mb, a fact worth fixing in our memory, as
different reasons favour the belief that it is precisely the M-type that
contains the very faintest stars. Its apparent velocity (_i.e._, the
proper motion) is so great that the star in 1000 years moves 3 deg., or as
much as 6 times the diameter of the moon. For this star, as well as for
its nearest neighbours in the table, observations differing only by a
year are sufficient for an approximate determination of the value of the
proper motion, for which in other cases many tens of years are
required.

Regarding the distribution of these stars in the sky we find that,
unlike the brightest stars, they are not concentrated along the Milky
Way. On the contrary we find only 6 in the galactic equator squares and
12 in the other squares. We shall not build up any conclusion on this
irregularity in the distribution, but supported by the general thesis of
the preceding paragraph we conclude only that these stars must be
relatively near us. This follows, indeed, directly from column 8, as not
less than eleven of these stars lie within one siriom