234 
ASTRONOMY: F. H. SEARES 
whence we conclude that the plate-errors have been successfully eliminated. 
The complete observation of a star, including the setting of the telescope, re- 
quires seven or eight minutes. 
Exposure-ratio — Components of 61 Cygni 
LOG Rl 
LOG R2 
DIFFERENCES 
f3 
9.39 
9 
50 
-0.11 
+0.03 
-0 
02 
+0 
05 
9.41 
9 
47 
-0.06 
+0.01 
+0 
01 
0 
00 
9.43 
9 
45 
-0.02 
-0.01 
+0 
03 
-0 
04 
9.39 
9 
50 
-0.11 
+0.03 
-0 
02 
+0 
05 
9.43 
9 
45 
—0.02 
—0.01 
+0 
03 
-0 
04 
9.42 
9 
48 
-0.06 
0.00 
0 
00 
0 
00 
9.42 
9 
47 
-0.05 
0.00 
+0 
01 
-0 
01 
9.42 
9 
47 
-0.05 
0.00 
+0 
01 
-0 
01 
9.44 
9 
45 
-0.01 
-0.02 
+0 
03 
-0 
05 
9.40 
9 
48 
-0.08 
+0.02 
0 
00 
+0 
02 
9.40 
9 
49 
-0.09 
+0.02 
-0 
01 
+0 
03 
9.39 
9 
50 
-0.11 
+0.03 
-0 
02 
+0 
05 
9.41 
9 
49 
-0.08 
+0.01 
-0 
01 
+0 
02 
9.39 
9 
49 
-0.10 
+0.03 
-0 
01 
+0 
04 
9.48 
9 
51 
-0.03 
-0.06 
-0 
03 
-0 
03 
9.44 
9 
50 
-0.06 
-0.02 
-0 
02 
0 
00 
Means 9.416 
9.481 
-0.065 
dbO.019 
±0.016 
±0.028 (log R) 
1.09 
1.27 
-0.18 
±0.053 
±0.045 
±0.078 (magnitude) 
Results are now available for about 150 stars whose spectral types and ab- 
solute magnitudes have been derived mainly from three sources: the exten- 
sive and valuable lists of spectroscopic parallaxes by Adams and Joy, which 
include types F, G, K, and M;^ the A stars known to belong to the Taurus 
Stream;^ and the B stars whose parallaxes have been determined by Kapteyn.^ 
The weight of the results is very unequal; some of the stars used as standards 
have been observed a dozen or more times, while for others only a single ob- 
servation is available. The nature of the correlation between color and lumi- 
nosity is nevertheless clearly enough indicated, although quantitatively it will 
be subject to considerable revision. 
The discussions thus far undertaken are very simple. The stars having abso- 
lute magnitudes between + 1 and — 1 were separated f om the others and ar- 
ranged according to spectral type. By plotting spectrum against exposure- 
ratio a curve was obtained (circles, fig. 1 ;) which shows the relation of these 
two factors for stars of zero absolute magnitude. The dwarf stars of the 
later spectral types, whose mean absolute magnitudes are 6 or fainter, were 
similarly treated, thus giving the second curve (points) shown in fig. 1. 
For M = 0 the variation is nearly linear up to K5; but beyond this point 
there seems to be no further increase in color with advancing spectrum, the 
latest K's (K6) and the various sub-classes of the M stars all having approxi- 
