SCIENCE 



[N. S. Vol. XXXIX. No. 992 



expect no better results than those obtained 

 with a filar micrometer. The best plan 

 may therefore prove to be to give a motion 

 to the wire nearly equal to that of the star, 

 whatever the declination of the latter, by a 

 suitable variation of the clockwork. The 

 best rate could readily be determined by ob- 

 serving stars at different distances from 

 the Pole. Successive settings should then 

 be made as with a filar micrometer, closing 

 the circuit on the chronograph only when 

 the bisection was satisfactory. A similar 

 setting should also be made for the declina- 

 tions. The two coordinates could thus be 

 determined with an accuracy substantially 

 the same as that of a filar micrometer. 

 Experience has shown that one star a min- 

 ute can be observed in both coordinates 

 with the transit micrometer. There can be 

 little doubt that positions could thus be ob- 

 tained with much greater accuracy than by 

 the methods now in use. The special ad- 

 vantage would be the elimination of system- 

 atic errors. 



A second method of determining posi- 

 tions, recently developed at the Allegheny 

 Observatory, is by plates taken with a 

 photographic doublet. Ordinary plates 

 must be replaced by those of plate glass. 

 By taking suitable precautions positions 

 may be determined of even the faintest 

 stars, with an accuracy at least equal to 

 that of a meridian circle. To obtain the 

 best results, the field should be about five de- 

 grees square on an 8 X 10 plate. The focal 

 length of the telescope would accordingly be 

 about two meters. The large field would 

 permit the constants of each plate to be de- 

 rived from stars as bright as the eighth 

 magnitude. The economy of this method 

 would be very great, as compared with a 

 meridian circle. The usefulness of the 

 latter instrument appears to be confined to 

 observations of the brighter stars. Ac- 

 cordingly, its aperture may be reduced. 



The ideal plan would apparently be to di- 

 vide the sky into regions five degrees 

 square and select in each, five or more 

 stars of about the eighth magnitude and of 

 approximately the same class of spectrum 

 as Class K, so that all should have about 

 the same color. The positions of these 

 should be determined with the greatest 

 possible accuracy with meridian circles, as 

 described above. Some brighter stars 

 should be included to render available the 

 vast number of observations of these ob- 

 jects made in the past. Positions of the 

 stars in the Gesellschaft Catalogues and 

 all fainter stars should be determined ) 

 photography. 



"Various attempts are now being made to 

 determine the absolute positions of the 

 stars by means of photography. It ap- 

 pears probable that a pier placed under 

 ground will remain free from irregular 

 motions, and that if this can be accom- 

 plished, the absolute positions of the stars 

 near the equator can be found by photog- 

 raphy. To determine the equinox, Venus 

 and Mercury should be photographed as 

 well as the sun. By the very satisfactory 

 cooperation of the Princeton, Tale and 

 Harvard observatories the position of the 

 moon is now determined by photography. 

 The results of a preliminary discussion in- 

 dicate an accuracy at least equal to that 

 of the best meridian determinations, those 

 of the Greenwich Observatory. 



Excellent progress is also being made in 

 determining the parallax of the stars by 

 photography. The recent increase in ac- 

 curacy is at least ten fold, or that of 

 another place of decimals. A hundredth 

 of a second of arc can now be determined 

 with greater accuracy than a tenth of a 

 second, twenty or thirty years ago. 



The just criticism has been made of 

 American astronomers that while they 

 have contributed more than their share of 



