April 25, 1913] 



SCIENCE 



647 



panion, however, is about one millionth that of the 

 atmosphere at the earth's surface. This greatly 

 surpasses the rarity of any other known eclipsing 

 star, but even then must be considered stellar 

 rather than nebular. The sun extended to Jupi- 

 ter 's orbit would not be so dense as this, and seen 

 from a neighboring star would certainly appear as 

 a stellar point. Except near the limb, the larger 

 component will be perfectly opaque. Through the 

 center of the star the extinction must be nearly 

 40 magnitudes, if the mass is talsen equal to the 

 sun 's mass. To allow for the translucence near 

 the beginning and end of eclipse, the relative 

 radii of the stars must be reduced still further, 

 which would tend to make the density of the 

 smaller star entirely normal without materially 

 diminishing the density of the large component. 

 The elements of the best orbit, which allowed for 

 darkening to zero at the limb, are as follows: 

 ratio of stellar radii, 0.10; inclination, 77°. 0; 

 radius of large star (distance of centers, unity), 

 0.387; light of each component, 0.50; relative sur- 

 face brightness, 0.01 ; hypothetical secondary min- 

 imum, 0"'.004; "equal -mass" densities, 2.4 X 10"' 

 and 2.4 X lO"". 



Harvard classes the spectrum as F^p. Luden- 

 dorif finds peculiarities in the radial velocity, and 

 is making a detailed spectroscopic study of the 

 system. 



Film Distortions on Small Fhotograj>hic Plates: 



F. E. Boss. 



By means of a reseau, tests were made of the 

 film distortions on the small photographic plates 

 used in the photographic zenith tube at Gaithers- 

 burg, Md. 



The cut Lumiere Sigma plates are 37 mm. long 

 by 27 mm. wide. ; in all forty plates were meas- 

 ured, of which 20 were dried in air in the usual 

 way, and 20 were immersed in alcohol before 

 drying. Distortions in only one coordinate, the 

 plates' length, were measured. The air-dried 

 plates showed large, irregular distortions. The 

 probable error of a measured 2 mm. space was 

 ±1.5ja; of a 22 mm. space, 9.6 ii. For larger 

 distances the probable error was somewhat smaller. 

 The corresponding probable errors of alcohol-dried 

 plates were 0.7 /i and 2.7 /i, respectively. The 

 probable error of an average distance on air-dried 

 plates was 6.5^; on alcohol-dried plates 1.4 |Cx. 



The maximum distortion observed on an air- 

 dried plate was 49 m ; on an alcohol-dried plate 

 10 /i. 



Air-dried plates always showed an expansion 



over about three fourths its distance from the 

 center. At this point the expansion, which 

 amounted to 19 /x, ceased, but the irregularities 

 became a maximum. 



No certain expansion of the alcohol-dried plates 

 could be detected. Air-dried plates which showed 

 unusually large distortions were resoaked Ln water, 

 dried in air and remeasured. The distortions were 

 found to have been notably changed, in both dis- 

 tribution and amount. They were again well 

 soaked and measured while wet. The distortions 

 were found to have disappeared. Other air-dried 

 plates showing large distortions were resoaked and 

 redried in alcohol. The distortions disappeared 

 in this case also. 



Some Effects of Sadiation upon Astronomical In- 

 struments: F. E. Boss. 

 Read, but no abstract was submitted. 

 Becent Progress in the Theory of the Galilean 

 Satellites of Jupiter: Kurt Laves. 

 The paper is a part of a report concerning the 

 Theory of Satellites in the Solar System, which 

 will appear in Vol. VI. of the ' ' Eueyclopaedie der 

 Math. Wissenschaf ten, ' ' published by the Academy 

 of Sciences of Gottingen. 



The modern investigations of de Sitter, Cookson, 

 Sampson and others were considered, and atten- 

 tion was drawn to the fact that the old Laplace- 

 Souillard theory is inadequate from a modern 

 standpoint. 



The Solar Botation in 1912: J. S. Plaskett. 



Two series of rotation plates were obtained in 

 1912, though with great diflS.eulty, owing to cloudy 

 and hazy weather. The first of these series, center 

 at X 5600, the special region allotted to Ottawa, 

 was obtained during June, and the second, center 

 at X 4250, the general region, during October. 

 The first series, consisting of 25 spectra at each 

 of the latitudes 0°, 15°, 30°, 45°, 60°, 75°, 80°, 

 85°, was measured and reduced by the writer, and 

 the velocity values obtained are well expressed by 

 a formula of the Faye type. 



V= (1.431 + 0.563 cos= 0) cos 0, 



where V is the linear velocity in kilometers per 

 second and <P the heliographic latitude. The coeffi- 

 cients of this formula lie between those of the two 

 series obtained in 1911 and the observed values 

 for the tyio years are in good agreement, the only 

 differences of any magnitude being at latitudes 

 45° and 75°, where the 1912 value is about 

 0.04 km. per second lower. The measurement and 

 reduction of the other series will likely give defi- 



