Febeuaby 31, 1896.] 



SCIENCE. 



285 



values of the astronomical constants were cor- 

 rected, and in the rigor with which the entire 

 discussion was carried through and the results 

 presented. 



A year or two after the appearance of Boss's 

 work, the new system for the Astronomische 

 ■Gesellschaft, constructed by Dr. Auwers, was 

 published. A very slight examination of this 

 work would show that its superiority to that of 

 Boss was at least open to question. The weak- 

 est point was that the proper motions depended ' 

 entirely on the observations of Bradley with 

 the old mural quadrant, which was known to 

 be subject to errors the amount of which did 

 ■not admit of determination. But this defect 

 did not prevent the general adoption of the 

 foreign system by American astronomers, even 

 in the case where the other would have been 

 most eminently appropriate, the official work of 

 "boundary surveys. 



There is one final and conclusive arbiter of 

 -all questions concerning the accuracy of pre- 

 dicted motions in the heavens. This arbiter 

 is subsequent observation. Let us wait a suffi- 

 cient length of time and see on which system 

 the positions of the stars are most accurately pre- 

 dicted. In certain features of the system and 

 in certain regions of the heavens the two works 

 differed so widely that a very few years of ac- 

 curate observations would suffice to settle the 

 question. 



About twenty years have elapsed since the 

 last observations on which either of the two 

 works was based. Within that time four cata- 

 logues of stars have appeared, founded on ob- 

 servations made at the respective observatories 

 of Pulkowa and Greenwich, prepared with all 

 the refinements of recent science, and therefore 

 superior to any before made. In these results, 

 combined with such conclusions as can be drawn 

 from the best previous observations, we have 

 the basis of a comparison which is found in the 

 number of the Astronomical Journal quoted in 

 the note found in the last number of Science. 

 Without going into technical details, it will suf- 

 fice to say that there are six separate and inde- 

 pendent features in which the respective sys- 

 tems difiered most largely. These six features, 

 tested by the four modern authorities just 

 quoted, showed the following average errors or 



difference between Boss's prediction and obser- 

 vations in different regions of the heavens, near 

 the epoch of 1880: 



It was then shown that, carrying back these 

 six special points of difference between the two 

 catalogues to the epoch of Bradley's observa- 

 tions, the actual differences between the two 

 were larger than any likely deviation of Boss 

 from the truth. In the most marked case the 

 difference consisted in ten discrepancies, all in 

 the same direction. Another very marked in- 

 stance occurs in a region of the heavens includ- 

 ing the northern part of the constellation An- 

 dromeda. In this region were found ten stars in 

 the A. G. catalogue. The Polkowa catalogue of 

 1895, the most carefully prepared that astron- 

 omy has yet had at its command, showed that 

 every one of these ten stars was in error in the 

 same direction, that direction being the same in 

 which they differed from the Boss system, and 

 by amounts which could not be reasonably at- 

 tributed to errors of the Pulkowa observations. 



One would suppose the conclusion so obvious 

 as to need no statement and admit of no ques- 

 tion. Fifteen years of the most refined obser- 

 vations show a continuing agreement of the Boss 

 system with observations which is most extra- 

 ordinary, and which cannot possibly be shared 

 by the other. This evidence, however, fails to 

 convince the writer of the note. He claims 

 that the results ' throw no new light on the sub- 

 ject.' If astronomers differ as to the question 

 whether the approach to perfect agreement 

 with observation above shown is conclusive, the 

 question would seem to be forever incapable of 

 decision. 



Again, in the case of ten separate stars in 

 which the deviations of the Bradley observa- 

 tions were all in the same direction, the writer 

 remarks: " So we can hardly escape the convic- 

 tion that our whole •conclusion may be vitiated 

 by a large error in a particular star. ' ' 



Here it would seem that the astronomers must 

 have recourse to legal advice to settle their dis- 

 pute. Only a member of the legal profession can 

 decide whether the concurrent evidence of ten 



