124 



SCIENCE. 



[N. S. Vol. XII. No. 291. 



as 0".5 excepting such as had already been 

 observed for parallax. 



It was hoped that among so large a num- 

 ber, nearly a hundred, some very near 

 neighbors should be found, but in case the 

 results should prove wholly negative it 

 would afford some satisfaction to know 

 that there are probably no more stars in 

 the northern skies within a certain distance 

 of us. 



This research was begun in the summer 

 of 1892, soon after Porter's Catalogue ap- 

 peared, and has been the problem of chief 

 attention on my part since that time. 

 There were 86 stars in my list and 13 in Dr. 

 Elkin's, of which I have completed the ob- 

 servations of 84 and Dr. Elkin 8. The orig- 

 inal plan was to observe each star on 

 three different nights near each of the two 

 epochs of maximum parallactic effect. For 

 each star when possible two suitable com- 

 panion stars were chosen on opposite sides 

 of the principal star and as nearly as pos- 

 sible at the same angular distance from it. 

 The observations were made in the custo- 

 mary symmetrical order SI, S2, S2, SI, SI 

 denoting the angular distance from one 

 companion star and *S2 the distance from 

 the other. 



At first it was intended to use the known 

 proper motions in the reductions, and it 

 was thought that three observations at each 

 of the two epochs would be sufficient to 

 show any parallax as great as 0".2, and any 

 such cases were to be further investigated. 

 Later it seemed to us to be desirable to 

 eliminate the effect of proper motion inde- 

 pendently, which can be quite thoroughly 

 accomplished by repeating the observations 

 through two more epochs in the reverse 

 order, and at the same time this enlarged 

 number of observations should furnish a 

 pretty fair approximation to the true value 

 of the parallax. 



The plan thus modified would give us 

 twelve complete observations for each star. 



which number was secured in nearly every 

 case. Each of these complete observations 

 furnishes an equation of condition of the 

 form : 



X -\- by + cs^ 11 



where x represents the required correction 

 to the assumed scale value, y the parallax, 

 s the correction to the assumed annual 

 proper motion, b the parallax factor de- 

 pending upon the positions of the stars 

 and that of the earth at the time, c = t — 

 1895.0 (1895.0 being about the middle of 

 the period covered by the observations) 

 and n equals the difference 81 — S2 minus 

 an assumed value for this difference. 



The normals from these equations of con- 

 dition have all been formed and a prelimi- 

 nary solution has only just been finished. 

 As to the results I may say that a little dis- 

 appointment was felt that no very large 

 parallaxes were found. However there were 

 two stars viz : 54 Piscium and Weisse 17" 

 322, which show a parallax of nearly 0".25 

 and which, therefore, if the results are con- 

 firmed by further observation, will place 

 them among the first ten or twelve nearest 

 stars so far as at present known. I have 

 selected for each of these stars two new 

 pairs of comparison stars and have nearly 

 completed a more extended series of obser- 

 vations of them. The final parallax will in 

 each case depend upon 56 complete obser- 

 vations instead of 12 as at present. 



A preliminary classification, according 

 to the magnitude of the parallax formed, 

 may be of some interest and is given in the 

 following table : 



Parallax. No. Stars. 



0".20 to 0".25 2 



.15 " .20 6 



.10 " .15 11 



.05 " .10 24 



.00 " .05 34 



—0 .05 " .00 8 



—0 .10 "—0 .05 5 



—0 .15 "—0 .10 2 



The probable error of a single observation 



