188 
ASTRONOMY: A. VAN MAANEN 
proper motion of O'MOO annually is only 0".018. It is clear that in such 
cases the probable error must not exceed a few thousandths of a second 
of arc if the results are to be valuable. 
In Contributions from the Mount Wilson Solar Observatory, No. 79, 
Adams and Kohlschiitter have given the absolute motions and absolute 
magnitudes of 100 stars with known parallaxes, for which they have 
determined the radial velocities. Nearly all these stars have large 
proper motion and are between the spectral types F and M. In the 
hope of completing this list by the addition of stars of the same spectral 
type, but of smaller proper motion, an investigation has been made to 
determine the accuracy with which parallaxes can be obtained with the 
60-inch reflector. Although in general the method employed was that 
used by Schlesinger and Slocum with the 40-inch Yerkes refractor, the 
following deviations from their procedure may be mentioned: 
The equivalent focal-length of the Cassegrain combination of mirrors 
employed is 80 feet, thus increasing the scale of the plates by one-quarter. 
The exposure time of 15 minutes gives stars as faint as magnitude 13; 
by avoiding the very faintest objects visible on the plates, we can still 
use stars between magnitude 9.5 and 12 for comparison purposes. The 
fainter stars must as a rule have smaller parallaxes, so that we have here 
the double advantage of using comparison stars with smaller parallaxes 
and with a smaller mean distance from the central star. 
The plates used are Seed 23, which have a fine grain; the star images 
therefore are sharper and the grain is less troublesome in the measuring. 
The plates are measured with the bHnk-arrangement of the stereocom- 
parator; two plates are compared directly without the use of any scale 
such as in ordinary measuring instruments. That the stereocompara- 
tor can, with some precautions, be used successfully for this kind of 
work was proved in a previous article {Astronomical Journal, 27, 140; 
1912). 
Special care has been taken to make the hour angles of the two plates 
to be compared nearly the same. The difference in hour angle has in 
no case been larger than 5°, while in most cases it was less than 2°. 
In comparing two plates directly we avoid in the final solution for 
the parallax the use of a third unknown quantity; this increases the 
weight of the resulting parallax. 
Finally, care has been taken to make the weight of the parallaxes 
resulting from a certain number of plates as large as possible "by dis- 
tributing the exposures as symmetrically as practicable within the year. 
There is, however, a difficulty sometimes affecting the use of the 60- 
inch reflector for this work. After a rapid change in temperature dur- 
