226 
ASTRONOMY: A. VAN MAANEN 
On the other hand, from the abundance of stars near the Trapezium we 
can conclude that most of them form a part of the system of the Orion nebula. 
From their connection with the nebulosity other stars can also be selected 
as probabje members of the group. Except for any internal motion, which, 
however, can be neglected in a first approximation, all these stars must have 
the same proper motion; they therefore enable us to drive any magnitude error 
in case such error does exist. 
Another advantage of a field hke that discussed here, is that, although the 
small field measured contains only two or three stars for which meridian ob- 
servations are available, we can now use for the reduction of our relative to 
absolute motion all the stars in the catalogues which seem to belong to the 
system of the Orion nebula. For this purpose twelve stars in Boss's Prelim- 
inary General Catalogue were found to be available. The motions previously 
derived were corrected in this way and reduced to absolute motions in Boss's 
system. The results for different groups of stars are as follows: 
13 stars which seem to be enveloped in nebulosity have 
M« = +0^52 zb oroo2o 
= -0!0024 ± 0.0024 
32 stars near the Trapezium have 
Ma= +0r0059 ± 0.0013 
^3 = -0.0021 =t 0.0010 
21 stars, found to be variable by different observers, 
iu« = +0.0046 ± 0r0009 
^5 = +orooii ± orooii 
Excluding here three stars for which the variability is uncertain, we find 
/z«= +0.0049 ± 0':0009 
^5 = -0^0003 ± oroolo 
From this it is clear that practically all these variables must belong to the 
system of the Orion nebula and that their variability is therefore due to a 
physical connection with the nebula. 
From the work of Buisson, Fabry and Bourget,^ we know there is some evi- 
dence that the nebula rotates about an axis NW — SE, the NE portion reced- 
ing, the SW approaching. As their measures are confined to the region 
within 2' from the Trapezium, I have investigated whether the stars meas- 
ured within the same area show the effect of such a rotation. As most of the 
stars seen in the densest part are likely to be on this side of the nebula, they 
ought to show a preponderance of motion from SW towards NE. The mean 
motion in this direction (27 stars) was found to be O'^OOIO zt 0W12, while the 
mean motion at right angles is OfOOOO =b O'^OOIS. This small systematic mo- 
tion is well within the limits of the probable error. 
For the background stars, that is, all stars not included in the other groups, 
the change in the mean values of //^ and m5 with magnitude, and the fair agree- 
