ASTRONOMY — RUSSSLL. lol 



and proper motions of all these stars have been determined. For 44 stars, 

 especially selected for investigation, the computations have been made by 

 least-squares. For the comparison stars, 243 in number, a shorter process, 

 of practically equal accuracy, was employed. 



The remaining 7 fields, of which 38 plates were obtained, could be photo- 

 graphed at only two epochs, owing to an accident to the color-screen. These 

 give the parallax of 8 bright stars (the principal objects in these fields), 

 whose proper motions are accurately known from meridian observations, 

 with almost the same accuracy as the other series. The whole number of 

 stars whose parallax has been determined is thus 295. 



The general discussion of the results is so far advanced that a summary 

 may here be given of the principal conclusions derived from the data for 

 the 287 stars which form the principal part of the work. 



(i) The results strongly support the theories of Newcomb and Kapteyn 

 concerning the distribution of the stars in space. Of the 44 parallax stars 

 (i. e., those especially picked out for investigation, mostly because of large 

 proper motion or some other indication of probable nearness), 31 give posi- 

 tive parallaxes, and only 3 negative ones, which exceed their probable errors, 

 while for the remaining 10 the computed parallaxes are less than their prob- 

 able errors. It is therefore evident that a large majority of these stars are 

 really nearer us than are the comparison stars, with reference to which their 

 parallax was determined, and that the criteria used in their selection are of 

 real value. Comparison of the individual results with Kapteyn's formulae 

 for the mean parallax of stars of determined proper motion and magnitude 

 confirms the latter in a very striking manner. 



The calculated parallaxes of the comparison stars, on the other hand, are 

 distributed in magnitude in very close accordance with the law of errors. 

 This confirms the conclusions of the above-mentioned astronomers that the 

 differences of parallax of stars of the eighth or ninth magnitude, taken at 

 random, are too small to be detected individually by the present methods of 

 observation. 



The computed proper motions of these stars, however, show, in addition 

 to the small errors of observation, unmistakable evidence of real proper 

 motion, whose average amount can be determined with considerable accuracy. 

 The corresponding parallax, on Kapteyn's theory, should be too small to be 

 detected. 



(2) The mean parallax of selected groups of the comparison stars may be 

 found with very high accuracy, and in this way much information may be 

 obtained. For example: 



(0) The means for stars near the center or near the edge of the plate, 

 and also those for stars of different magnitudes, do not differ sensibly. 



