December 16, 1910] 



SCIENCE 



879 



This has been made by Mr. W. B. Blaikie, of 

 Edinburgh; it is very simple and ingenious, and 

 is useful when one is only working to half degrees. 

 He has lithographed on two sheets of celluloid 

 stereographic projections of parallels and merid- 

 ians, the poles in each case being on the circum- 

 ference. One of these sheets is fi.xed and the 

 second, which is placed above it, can be turned 

 about a pin at the center. In this way five parts 

 of the triangle can be read off at once. 



An Interesting Speotroscopio Binary, OG Her- 



culis: S. A. Mitchell. 



While at the Yerkes Observatory in the earlier 

 part of this summer, a spectroscopic binary of 

 more than usual interest was found in 96 Her- 

 culis, a star of 5.1 magnitude, 15 type, with 

 many good lines. The first plate measured showed 

 the lines to be triple and the ternary character 

 of the star has been abundantly verified. The 

 results of measures from two plates are as fol- 

 lows: 



Plate taken on June 6, 1910. 



Component I., radial velocity = — 26 km. from 

 18 lines. 



Component II., radial velocity ^ + 25 km. from 

 6 lines. 



Component III., radial velocity = -|- 62 km. 

 from 9 lines. 



Component I. is very strong, II. and III. are 

 sharp but faint. 



Plate taken on June 24, 1910. 



Component A, radial velocity = — 85 km. from 

 8 lines. 



Component B, radial velocity = — 19 km. from 

 16 lines. 



Component C, radial velocity = + 33 km. from 

 8 lines. 



Components A and C are faint, but sharp; B is 

 strong. 



Changes are rapid, the period (if such can be 

 said to exist in a three-body system) is a few 

 days. 



On the Accuracy of the Star Positions of -the 



Harvard Sky: H. H. Tubneb. 



The Harvard Sky consists of fifty-five plates, 

 each covering about thirty degrees square. The 

 scale is rather less than one tenth that of the 

 Astrographic Catalogue, a r^seau interval of five 

 millimeters corresponding to about fifty-two min- 

 utes at the center of the field. The plates are 

 therefore not intended for giving accurate posi- 

 tions, but It is convenient to know what kind of 



accuracy is obtainable from the plates, and fur- 

 ther it is of interest to know the optical distor- 

 tion of a lens covering such a wide field. Two 

 plates have been partially measured with some 

 care, the measures being confined to the neighbor- 

 hood of rectangular axes through the plate center. 

 The first plate (R. A. 6" 0'"; Decl. 0°) was meas- 

 ured some years ago and an outward optical dis- 

 tortion of 0".024 r" was indicated, r being the dis- 

 tance from the plate center expressed in rfiseau 

 intervals. The second plate was measured re- 

 cently by Mr. G. H. Hamilton, who used a much 

 larger number of stars than were used before. 

 The value for the distortion was found quite inde- 

 pendently to be 0".030 r^ in one coordinate and 

 0".036 r^ in the other. The uncertainty of the 

 determination is due to the confusion with the 

 scale value: for the differences may be written 

 (say) it O".O06 r(7-= — 10"); which does not ex- 

 ceed 2".3 between r^O and r=VJ. The point 

 on the plate from which this distortion radiates 

 may be approximately identified as follows: Let 

 the coordinates be (a, 6) so that the displacement 

 in X is k(a> — a) S(x — a)- + (y — hy\. The 

 terms in se' and xy can arise from tilt of plate; 

 but the term — hay- can not arise from tilt and 

 enables us to find a. Similarly 6 can be found 

 from the displacement in y. When this distortion 

 and a slight tilt of the plates are allowed for, the 

 resulting star measures agree closely with the 

 calculated measures; which suggests that if large 

 plates could be made sufficiently flat, large fields 

 might be photographed with accuracy. This ob- 

 ject might be attained as follows: Let a large 

 glass surface be carefully planed and ruled with 

 rSseau lines; and placed in the focus of a wide 

 angle lens. Several small plates could be placed 

 film down on this surface, to receive the images 

 of the stars and of the portion of the rfiseau 

 which they covered. They could be developed and 

 measured as separate plates, all accurately con- 

 nected by the rgseau lines of the matrix. 



The Photometric Magnitude of Eros in 1903: S. 



I. Bailey. 



The variability in light of Eros was announced 

 by E. von Oppolzer in 1901. At this time the 

 range of variation was said to be two magnitudes. 

 Observations somewhat later at the Harvard Ob- 

 servatory-, by Professor Wendell, showed that at 

 times the range of variation was very small or 

 entirely lacking. At the suggestion of the di- 

 rector of Han-ard Observatory, the writer under- 

 took an extensive study of the changes in light 



