26 



SCIENCE 



[N. S. Vol. XXXVII. No. 94» 



of plates showing trails of the Pleiades. These 

 plates were exposed on several different nights. 

 Proper precautions were taken to leave the tele- 

 scope as undisturbed as possible. The trails were 

 measured at the Allegheny Observatory by setting 

 the micrometer wire on the mean position of suc- 

 cessive portions of the trail, each corresponding to 

 an interval of about three seconds of time. It 

 was found that on each evening the trails con- 

 clusively show the presence of irregular refrac- 

 tions, portions of the trails remaining on one side 

 (of the mean position for that plate) sometimes 

 nearly a minute. On two of the plates simultane- 

 ous trails of Alcyone and Merope were measured 

 and these show a remarkable parallelism. The 

 result of this experiment tends strongly to con- 

 firm the deductions of Nusl and Fric. 

 The OrUt of B Canis Majoris: Frank C. Jordan. 

 This star, one of the few Algol variables with 

 an advanced type spectrum, F, was discovered to 

 be a light variable, by Sawyer in 1887, and to 

 have variable radial velocity, by Frost in 1905. 

 Forty-nine plates have been obtained at Allegheny 

 in five seasons of observation. The light period 

 has been very accurately determined and is taken 

 as the definitive orbital period. 



The elements are as follows: P = 1.13595 

 days; e = 0.138; w = 195°.86; ir = 28.64 km.; 

 7 = — 43.45 km.; 7 = 1908, Jan. 25.296, Light 

 Minimum; a sin i = 453,500 km.; m'^/{m + m')- 

 = 0.00288 O- 



Accurate photometric observations combined with 

 the orbital elements, and the fact that the spec- 

 trum is much like that of our sun, will give an 

 independent means of finding the star's parallax. 

 New Eclipsing Variable Stars: Joel Stebbins. 



It is probable that a considerable number of 

 the short-period spectroscopic binaries are eclipsing 

 variable stars, and the writer reports progress on 

 tests of these systems with the selenium photom- 

 eter. It seems practically certain that a Virgin! s 

 and a Coronse Borealis are such variables, and 

 /S Scorpii is suspected. The first, Spica, has a 

 period of 4.014 days, and shows two minima, a 

 primary of something greater than 0.10 mag., and 

 a secondary of about 0.08 mag. The main eclipse 

 of a CoronEB was observed only once, but it was 

 a decrease of at least 0.12 mag., and came at the 

 predicted time. This star is a member of the 

 extended Ursa Major group, and has a period of 

 17.36 days. The suspicions of the variation of 

 /3 Scorpii seem well founded, but need confirma- 

 tion. 



The results emphasize the urgent need of fur- 

 ther tests on similar stars. 



The Scale of the Durchmusierung : E. C. Pick- 

 ering. 



The number of stars in the Bonn Durchmus- 

 terung is about 460,000; in tlie Cordoba, 490,000; 

 in the Cape, 450,000; total, 1,400,000, or, omitting 

 duplicates, 1,100,000 stars. In the Harvard An- 

 nals, 23, the deviations, from the photometric 

 scales, of the magnitudes 9.0 and brighter are 

 discussed. Volumes 70 and 74 permit a similar 

 study of the fainter stars and of those in the 

 Cordoba Durchmusterung whose average bright- 

 ness differs as much as a magnitude in different 

 hours, the brightness increasing as the number of 

 stars per degree increases. The photometric mag- 

 nitude of Bonn 9.5 is 10.5; of Cordoba 10 it is 

 12.0 in zones — 22° to — 42°, and 11.7 in zones 



— 43° to — 52°. Stars of a given Durchmuster- 

 ung magnitude are half a magnitude brighter in 

 zones ■ — 43° to ■ — 52° than those in the zones 



— 22° to —42°. 



The deviations of the Cape Durchmusterung 

 magnitudes from the photometric magnitudes are 

 reduced one half by grouping stars on the same 

 plate and applying a correction for the class of 

 spectrum. The average deviation in a provisional 

 reduction of 12 plates is only ± 0.2 mag. 

 The Progress of the Revised Draper Catalogue: 



Annie J. Cannon. 



The purpose of this catalogue is to give the 

 class of spectrum of all stars down to about the 

 eighth magnitude, distributed over the entire sky. 

 Photographs having exposures of sixty minutes 

 have been made at Cambridge and Arequipa with 

 a prism of small angle placed before the objective- 

 of 8-inch doublets, and all stars will be classified 

 which are bright enough to be seen clearly and 

 have not already been published in H. A., 28,. 

 and S. A., 56, 4 and 5. Besides the classifica- 

 tion the catalogue will give the photographic mag- 

 nitude, the identification in some position cata- 

 logue, and the place for 1900. The catalogue was 

 started in October, 1911. "With an hour's observa- 

 tion each day, about 5,000 spectra can be classified 

 in a month. This rate has been maintained since 

 February, 1912, so that at the beginning of 

 August, 45,000 stars have been classified. Charts 

 were exhibited showing that about one third of 

 the sky has been covered. 



A Design for a New Form of Spectrograph Col- 

 limator: Philip Fox. 

 This form was proposed for use with the small 



