ASTRONOMY: R. G. AITKEN 
533 
Table III shows the relation between magnitude and angular sepa- 
ration. Since the progressive increase in the number of double stars 
as the angular distance diminishes is shown by stars of every magni- 
tude class, it cannot be explained as merely a perspective effect, the 
more distant pairs appearing at smaller angles, for according to our 
present conceptions, stars of a given magnitude (e.g., from 8.6 to 9.0) 
are approximately at the same average distance from us. The relative 
angular separations of the double stars of this magnitude therefore 
reflect the relative linear distances between the components and it 
follows that visual double stars as a rule revolve in relatively small orbits. 
The spectral classification of the majority of the 5400 stars under 
discussion is at present unknown. I have compared the list with the 
Draper Catalogue in volume 50 of the Harvard Annals, thus securing 
TABLE III 
The Distribution of Double Stars in the Northern Hemisphere by Angular 
Distance and Magnitude 
O'OO 
0'51 
(0"00) 
I'Ol 
2*01 
3*01 
4.01 
5roi 
DIST. 
TO 
TO 
TO 
TO 
TO 
TO 
TO 
AND 
0.50 
1.00 
a. 00) 
2.00 
3.00 
4.00 
5.00 
OVER 
Mag. to 6.5 
75 
63 
(138) 
83 
62 
41 
31 
99 
6.6 to 7.0 
82 
52 
(134) 
59 
42 
40 
21 
14 
7.1 to 7.5 
103 
67 
(170) 
99 
64 
48 
31 
29 
7.6to8.0 
178 
132 
(310) 
164 
107 
85 
63 
26 
8.1 to 8.5 
310 
223 
(533) 
285 
173 
128 
111 
21 
8.6to9.0 
508 
413 
(931) 
532 
317 
217 
191 
11 
Total 
1256 
954 
(2206) 
1222 
765 
559. 
448 
200 
the spectral data for the brighter stars, and Prof. E. C. Pickering and 
Miss Annie J. Cannon have been kind enough to compare the list of 
double stars north of + 60° and that of the double stars in the first two 
hours of Right Ascension v/ith the new Draper Catalogue now in course 
of preparation at Harvard, which is designed to extend to 9.0 magnitude. 
Combining the data from these three sources and excluding duplicates 
we have Table IV. 
The distribution of these stars with respect to the galactic plane 
closely resembles that of the stars in general; Classes B and A crowd 
tow^ard that plane, classes F, G and K show a more uniform distribution. 
The totals show how marked is the preference of the visual double stars 
for Classes A and F. It may be added that 46 of the 83 Class B stars 
belong to the subclasses B8 and B9 and that 104 of the 124 Class K stars 
are marked 'K' or 'KO.' Clearly, close visual double stars are rare among 
stars of either very early or very late spectral class. This confirms my con- 
