OCTOBKK 7, 1898.] 



SCIENCE. 



453 



clusters remarkable for the number of 

 variables are Messier 5, Omega Centauri 

 and Messier 15. These four clusters to- 

 gether have 393 variables. At the present 

 time the periods and light curves of more 

 than one hundred have been determined. 

 These periods are generally less than 24 

 hours, and are characterized by an abrupt 

 increase in light and a slower decrease. 

 These results have all been obtained by 

 photography. They would have been im- 

 possible with instruments of the same size 

 by visual methods. The photographs were 

 made with a 13" telescope. The faintest 

 variables are quite invisible in the same 

 telescope, and it is doubtful if they can be 

 well studied visually in a telescope of less 

 than 30" or 40" aperture. The shortest 

 period of these variables is 6" 11™, which 

 makes it the shortest period known. 



Professor Bailey's remarkable work was 

 rendered extremely interesting by the ex- 

 hibition of some of the original photo- 

 graphs and of the numerous charts of the 

 change in brightness of these variables 

 which he had been able to prepare. Pro- 

 fessor E. E. Barnard stated that he was sur- 

 prised that this great discovery had at- 

 tracted so little attention. He knew the 

 photographic plate could be depended on. 

 The great Yerkes telescope very readily 

 verified the periods. He had verified three 

 of Professor Bailey's periods, but was afraid 

 the visual was inferior to the photographic 

 method. Father Hagen stated that the 

 stars were on the limit of visibility in the 

 12-inch telescope, and there seemed to be 

 no charts of these clusters. Professor Pick- 

 ering stated that photographic charts of the 

 ■clusters would be furnished by the Harvard 

 'College Observatory. 



Mr. Albert S. Flint, in a contribution on 

 ' Meridian Observations for Stellar Paral- 

 lax,' presented a summary of the results of 

 the observations for stellar parallax, made 

 at the Washburn Observatory, on a list of 



nearly one hundred stars. The method 

 employed is that of difierences of meridian 

 transits. The value of the method is con- 

 firmed by the good agreement with other 

 accurate determinations of stellar parallax 

 and by the absence of any plainly indicated 

 systematic errors. The first comparison 

 gave an average difference of only 0".03 on 

 the fifty stars elsewhere observed. The 

 probable error of the definitive parallax for 

 any one star varies from 0".030 to 0".040, 

 except for great zenith distances. The 

 values depend upon from forty to sixty ob- 

 servations extended through from four to 

 six epochs. A new series of similar obser- 

 vations has been begun, which will include 

 second-magnitude stars and binaries. The 

 new Eepsold transit micrometer will be em- 

 ployed, with which the observer main- 

 tains bisection with a movable thread, 

 ■while the chronograph signals are made 

 automatically by the revolving micrometer 

 head. 



Professor Comstock said that while the 

 average accuracy of the method is not as 

 great as with the heliometer or by photog- 

 raphy more observations can be made and 

 gross error eliminated. The method is use- 

 ful for a general parallax survey. 



Professor Charles Lane Poor, in a paper 

 on ' The Direct Grating Spectroscope and 

 its Applications to Stellar Photography,' 

 stated that already in 1891 and 1892 he 

 had experimented with the direct vision 

 grating in photographing star spectra, and 

 had taken the problem up afresh in the 

 fall of 1897, using both a 2-inch grating 

 and a specially ruled grating of 6.5 inches 

 aperture and 39 inches radius of curvature. 

 Properly mounted on a guiding telescope 

 good photographs of stars to fourth magni- 

 tude resulted. Many of these were exhib- 

 ited by Dr. Poor. Aside from the ad- 

 vantage of securing normal spectra, the 

 method permits the determination of the 

 relative wave-length shift by photograph- 



